How to Use this Volume This volume is divided into the following sections (see index in the following page; click on the corresponding section name to go to the section): • Program. Full description of the program including, in chronological order, the information about the sessions: Name, 4 character code, room, chairperson and titles and authors of each communication. Each communication is linked to the corresponding abstract in the Abstracts section. • Abstracts. Contains the abstracts of all the communications presented at the conference, alphabetically ordered by titles. Each abstract contains the title, authors and affiliations, abstract and keywords. • List of keywords. All the keywords used in the different abstracts are listed in alphabetical order. For each keyword, we show the pages of the abstracts where this keywords appear. Each page is a link to the corresponding abstract. • List of authors. All authors and coauthors of communications are listed in alphabetical order. For each abstract authored by a person, we display the 4 character code of the corresponding session and the page of the corresponding abstract (linked to the abstract in the Abstracts section). The code of the session appears in boldface if the author is presenting the communication. 1 Contents Program 2 Abstracts 45 List of Keywords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378 List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387 1 Program Monday - 17th June Opening Session Room A (Paraninfo) - 9:15 Session MO1A (plenary) Plenary 1 Chaired by Stan Zionts Room A (Paraninfo) - 10:00 1. Creating, Structuring, and Using Multiple Objectives [10:00 - 11:00] by Ralph Keeney (*). Session MO2B (contributed) Reference Point Methods Chaired by Willem K. M. Brauers Room B (Milton Friedman) - 11:20 1. On the Optimal Inputs Balance in Terms of the Educational Multiobjective Model: the Case of Primary and Secondary Education Students [11:20 - 11:45] ´ by Mariano Luque (*), Luis Alejandro L´opez Agudo, Oscar Marcenaro-Guti´errez. 2. Synhronous Usage of Parameterized Achievement Scalarizing Functions in Interactive Compromise Programming [11:45 - 12:10] by Yury Nikulin (*), Volha Karelkina. 3. Goal Programming Model for Management Accounting: A New Typology [12:10 12:35] by Sheila McGillis (*), Belaid Aouni. 4. Optimization of Investment in Shares by Multi-Objective Optimization [12:35 13:00] by Willem K. M. Brauers (*). 2 Session MO2C (contributed) Room C (Paul A. Samuelson) - 11:20 Interactive Multi-Objective Optimization (I) Chaired by Murat Koksalan 1. An Interactive Algorithm for Multiobjective Mixed Integer Programming Problems [11:20 - 11:45] by Ozgur Ozpeynirci (*), Murat Koksalan, Banu Lokman. 2. Solving a Mutiobjective Dynamic Problem with a Stochastic Transition Function and Stochastic Objective Function by Interactive Procedure [11:45 - 12:10] by Maciej Nowak (*), Tadeusz Trzaskalik. 3. An Enumerative Cutting Plane Approach to Integer Linear Vector Optimization Problems [12:10 - 12:35] by Walter Habenicht (*). 4. An Interactive Algorithm for Multi-objective Integer Programs [12:35 - 13:00] by Murat Koksalan (*), Pekka Korhonen, Banu Lokman, Jyrki Wallenius. Session MO2D (contributed) Room D (Franco Modigliani) - 11:20 Environmental Multi-Criteria Decision Making (I) Chaired by Judit Lienert 1. MCDM and GIS to Identify Land Suitability for Agriculture [11:20 - 11:45] by Mendas Abdelkader (*). 2. Problems, Objectives and Ecosystem Services in Protected Areas: A Mediterranean Case Study [11:45 - 12:10] by Marina Segura (*), M´onica de Castro Pardo, Concepcion Ginestar, Concepci´on Maroto. 3. Constructing Ecological Value Functions for River Rehabilitation: Expert Elicitation with Different Aggregation Schemes [12:10 - 12:35] by Judit Lienert (*), Simone Langhans, Peter Reichert. 4. Abstract REMOVED [12:35 - 13:00] 3 Session MO2E (contributed) Goal Programming Chaired by Belaid Aouni Room E (James M. Buchanan) - 11:20 1. Goal Programming Approach to Workforce Scheduling Problem for a Restaurant [11:20 - 11:45] by G¨ ul¸cin C ¸ ¨ol (*), Servet Hasg¨ ul. 2. Error Correction Model in Classification by Using Multiple-Criteria and MultipleConstraint Levels Linear Programming [11:45 - 12:10] by Bo Wang (*), Yong Shi, Wei Xianhua. 3. Augmented Goal Programming with Multi Fuzzy Targets and Coefficients [12:10 12:35] by Hocine Mouslim (*), Mustapha Belmokaddem, Sakina Melloul. 4. Some Epistemological Considerations of the Goal Programming Model [12:35 13:00] by Belaid Aouni (*). Session MO2F (invited) Room F (Lawrence R. Klein) - 11:20 Evolutionary Multiobjective Optimization (EMO-I) Chaired by Dimo Brockhoff 1. Multi-Objective Reinforcement Learning using Sets of Pareto Dominating Policies [11:20 - 11:45] by Kristof Van Moffaert (*), Madalina M. Drugan, Ann Now´e. 2. Adaptive Guided Evolutionary Multi-Objective Optimization [11:45 - 12:10] by Florian Siegmund (*), Kalyanmoy Deb, Amos Ng. 3. FEMOEA: a Fast and Efficient Multi-Objective Evolutionary Algorithm [12:10 12:35] by Juana L´opez-Redondo (*), Jos´e Fern´andez-Hern´andez, Pilar Mart´ınez-Ortigosa. 4. Dynamic Resampling for Guided Evolutionary Multi-Objective Optimization of Stochastic Systems [12:35 - 13:00] by Florian Siegmund (*), Kalyanmoy Deb, Alexander Karlsson, Amos Ng. 4 Session MO2G (invited) Room G (Robert M. Solow) - 11:20 Socially Responsible Investment Decision Making (I) Chaired by Blanca P´erez-Gladish 1. A Multi-dimensional Corporate Social Performance Versus a Multi-dimensional Corporate Social Irresponsibility [11:20 - 11:45] by Homayoon Shalchian (*), Kais Bouslah, Khalid El badraoui, Jean-Jacques Lilti, Bouchra M’Zali. 2. Markowitz’s Investment Problem under Multicriteria, Uncertainty and Risk [11:45 - 12:10] by Vladimir Korotkov (*), Vladimir Emelichev. 3. Tri-criterion Inverse Portfolio Optimization with Application to Socially Responsible Mutual Funds [12:10 - 12:35] by Maximilian Wimmer (*), Markus Hirschberger, Ralph E. Steuer, Sebastian Utz. 4. Synthetic Indicators of Mutual Funds’ Environmental Responsibility: An application of the Reference Point Method [12:35 - 13:00] by Blanca P´erez-Gladish (*), Jose Manuel Cabello, Paz M´endez Rodr´ıguez, Francisco Ruiz. Session MO3B (contributed) AHP/ANP: Methodological Issues (I) Chaired by Ludmil Mikhailov Room B (Milton Friedman) - 14:30 1. Multi-Criteria Decision Making Applications in Higher Open and Distance Learning Systems [14:30 - 14:55] by Zehra Kamisli Ozturk (*). 2. A Decision Model for Identifying and Prioritizing the Capability Gaps in Defense Planning [14:55 - 15:20] by Ahmet Kandakoglu (*), Y.Riza Kahraman, Ilker Topcu. 3. Microsoft Excel as a Tool for Solving Multicriteria Decision Problems [15:20 - 15:45] by Radomir Perzina (*), Jaroslav Ramik. 4. Multiobjective Evolutionary Approach to Preference Elicitation in the Analytic Hierarchy Process [15:45 - 16:10] by Ludmil Mikhailov (*). 5 Session MO3C (contributed) Multi-Attribute Utility Theory (I) Chaired by Hannele Wallenius Room C (Paul A. Samuelson) - 14:30 1. Grey Numbers and Multi-Criteria Decision Making [14:30 - 14:55] by Alan Pearman (*), Naiming Xie. 2. Multicriteria Cognitive Map: an Integrated Tool for Building a Multicriteria Evaluation Model [14:55 - 15:20] by Teresa Cipriano Rodrigues (*), M´onica Duarte Oliveira. 3. Development of Interactive Support System Using Spreadsheet based on Multiattribute Utility Analysis [15:20 - 15:45] by Hayashida Tomohiro (*), Nishizaki Ichiro, Imai Mitsuhiro. 4. Multi-Attribute Online Reverse Auctions [15:45 - 16:10] by Hannele Wallenius (*), Pham Long, Jeffrey Teich, Jyrki Wallenius. Session MO3D (contributed) Room D (Franco Modigliani) - 14:30 Agricultural & Forestry Resources Management Chaired by Carlos Romero 1. A Model for a Spatial Forest Problem Taking into Account Economic and Environmental Objectives [14:30 - 14:55] by Monica Hernandez (*), Rafael Caballero Fern´andez, Trinidad G´omez, M. Amparo Leon, Juli´an Molina Luque. 2. Quantifying Public Preferences for Evaluation Multifunctional Agriculture System, using Analytic Hierarchy Process [14:55 - 15:20] by Inmaculada Marqu´es (*), Jos´e Lu´ıs P´erez-Salas, Baldomero Segura Garc´ıa del R´ıo. 3. Hydrological Constraints in Optimal Timber Harvest Scheduling Problems: A Case Study in Eucalyptus Plantations [15:20 - 15:45] by Juan Carlos Gim´enez (*), Mercedes Bertomeu Garc´ıa, Luis D´ıaz-Balteiro. 4. Multiple Criteria Methods in Decision Support Systems for Forest Management [15:45 - 16:10] by Marina Segura (*), Concepci´on Maroto, Duncan Ray. 6 Session MO3E (contributed) Advancements in MCDM Theory Chaired by Ralph E. Steuer Room E (James M. Buchanan) - 14:30 1. Biases and Path Dependency in the Even Swaps Method [14:30 - 14:55] by Raimo H¨am¨al¨ainen (*), Tuomas Lahtinen. 2. NIMBUS in Interactive Multiobjective Optimization under Uncertainty [14:55 15:20] by Kaisa Miettinen (*), Jyri Mustajoki, Theodor Stewart. 3. Evidential Reasoning Rule for Criteria Combination in MCDM [15:20 - 15:45] by Ling Xu (*), Jian-Bo Yang. 4. A Careful Look at the Importance of Criteria and Weights [15:45 - 16:10] ¨ orni, Kari Silvennoinen. by Jyrki Wallenius (*), Pekka Korhonen, Anssi O¨ Session MO3F (contributed) Preference Based EMO Chaired by Ignacy Kaliszewski Room F (Lawrence R. Klein) - 14:30 1. The Global WASF-GA: An Evolutionary Algorithm in Multiobjective Optimization to Approximate the whole Pareto Optimal Front [14:30 - 14:55] by Ana Bel´en Ruiz (*), Mariano Luque, Rub´en Saborido-Infantes. 2. Theoretically Analyzing Simple Interactive Evolutionary Multiobjective Optimizers [14:55 - 15:20] by Dimo Brockhoff (*). 3. Optimization of the Operation of the Auxiliary Services of Power Plants using Preference-based Evolutionary Optimization [15:20 - 15:45] by Ana Bel´en Ruiz (*), Jose Manuel Cabello, Mariano Luque, Francisco Ruiz. 4. Framing Mechanical Design as Multiple Criteria Decision Making Processes [15:45 - 16:10] by Ignacy Kaliszewski (*), Tomasz Kiczkowiak, Janusz Miroforidis. 7 Session MO3G (invited) Value-Focused Thinking Chaired by Johannes Siebert Room G (Robert M. Solow) - 14:30 1. Creating a Balanced Scorecard by using Value-Focused Thinking [14:30 - 14:55] by Joschka Muetterlein (*), Reinhard Kunz, Johannes Siebert. 2. Facing the Energy Transition: An Illustration how Value-Focused Thinking helps to Find the Right Answers [14:55 - 15:20] by Peter Krampf (*), Johannes Siebert. 3. Generating an Optimal Set of Alternatives using Value-Focused Thinking [15:20 15:45] by Johannes Siebert (*). 4. Concept and Empirical Verification of the Benefit of Value-Focused Thinking [15:45 - 16:10] by Reinhard Kunz (*), Johannes Siebert. Session MO4B (invited) Room B (Milton Friedman) - 16:30 New Methods for Multi-Objective Optimization (I) Chaired by Kaisa Miettinen 1. An Interactive Algorithm for Bi-objective Routing Problems [16:30 - 16:55] by Diclehan Tezcaner (*), Murat Koksalan. 2. Multi-Objective Modelling and Simulation in New Product Development (NPD) using the Belief Rule-Based (BRB) Methodology [16:55 - 17:20] by Emanuel Savan (*), Yu-Wang Chen, Ling Xu, Jian-Bo Yang. 3. A Conic Scalarization Method in Multi-Objective Optimization [17:20 - 17:45] by Refail Kasimbeyli (*). 4. Nonconvex Pareto Navigator for Interactive Multiobjective Optimization [17:45 18:10] by Markus Hartikainen (*), Kathrin Klamroth, Kaisa Miettinen. 8 Session MO4C (invited) Room C (Paul A. Samuelson) - 16:30 Approximate Solutions and Optimality Conditions in Vector Optimization Chaired by Vicente Novo 1. Approximations of Set-Valued Optimization Problems [16:30 - 16:55] by Rub´en L´opez Montoya (*), Elvira Hern´andez. 2. Efficient Solutions for Non-Regular Multiobjective Problems [16:55 - 17:20] by Mar´ıa Beatriz Hern´andez Jim´enez (*), Rafaela Osuna-G´omez, Gabriel RuizGarz´on. 3. Variational-Like Inequality Problems as an Important Tool to Solve Vectorial Optimization Problems [17:20 - 17:45] by Gabriel Ruiz-Garz´on (*), Lucelina Batista dos Santos, Rafaela Osuna-G´omez, Antonio Rufi´an-Lizana. 4. Characterization and Properties of Approximate Proper Solutions in Vector Optimization Problems [17:45 - 18:10] by Lidia Huerga (*), C´esar Guti´errez, Bienvenido Jim´enez, Vicente Novo. Session MO4D (invited) Room D (Franco Modigliani) - 16:30 Methods with Guaranteed Performance for Multi-Objective Combinatorial Optimization (I) Chaired by Xavier Gandibleux 1. A Parameterized Approximation for Matroids with Multiple Objectives [16:30 16:55] by Lydia Tlilane (*), Laurent Gourv`es, J´erˆome Monnot. 2. A Primal-Dual Approach for Optimizing Ordered Weighted Average in Perfect Matching Problems [16:55 - 17:20] by Paul Weng (*), Viet Hung Nguyen. 3. Approximations for the Multi-objective Spanning Tree Problem [17:20 - 17:45] by Anisse Ismaili (*). 4. Surrogate-Based Algorithm for Computing an Upper Bound set for the 0/1 BiObjective Bi-Dimensional Knapsack Problem [17:45 - 18:10] by Audrey Cerqueus (*), Xavier Gandibleux, Anthony Przybylski. 9 Session MO4E (Wiley Prize) Wiley Practice Prize Competition Chaired by Theodor Stewart Room E (James M. Buchanan) - 16:30 1. Application of an Intelligent Decision System to Assess Students’ Academic Performance in University Early Admission [16:30 - 16:55] by Yu-Wang Chen (*), Yue Chen, Ling Xu, Jian-Bo Yang. 2. Triple Bottomline, Hyper-radial-Visualisation-Based ’Decision-Making by Shopping’ for a Land Use Management Problem using Evolutionary Multi-objective Optimisation [16:55 - 17:20] by Oliver Chikumbo (*), Kalyanmoy Deb, Erik Goodman. 3. Integrating Qualitative Assessment and Fuzzy AHP to Prioritize Investments in an Energy Efficiency Program of Favelas in the City of Rio de Janeiro [17:20 - 17:45] by Jose Roberto Ribas (*), Mariana Rocha, Juliana Ribas Severo. 4. Multi-Objective Dynamic Scheduling Optimization Model for Real-Time Rebalancing Control of the Hybrid Assembly Lines [17:45 - 18:10] by Wai Keung Wong (*). Session MO4F (invited) Room F (Lawrence R. Klein) - 16:30 Multi-Criteria Models applied to Sustainable Environment and Efficient Use of Energy Chaired by Carlos Enrique Escobar-Toledo 1. Multicriteria Models Applied to Sustainable Enviroment and Efficient Use of Energy [16:30 - 16:55] by Lol-chen Alegr´ıa Mej´ıa (*), Carlos Enrique Escobar-Toledo. 2. Elicitation Procedures for the Comparison of Decisional Maps [16:55 - 17:20] by Valerie Brison (*), Marc Pirlot. 3. A Stakeholder based Multi-Criteria Evaluation Framework for City Distribution [17:20 - 17:45] by Lauriane Milan (*), Cathy Macharis, Sara Verlinde. 4. Multi-Criteria Decision Making: Exergy and Energy Life Cycle Analyses for Plastic Individual Disposable Bags versus other Material [17:45 - 18:10] by Carlos Enrique Escobar-Toledo (*), Lol-chen Alegr´ıa Mej´ıa, B´arbara Ram´ırez. 10 Session MO4G (contributed) Risk MCDM Modeling (I) Chaired by Rafael Sarkisyan Room G (Robert M. Solow) - 16:30 1. Multi-Criteria Decision-Making Perspective on the Privacy versus Security Conundrum [16:30 - 16:55] by Sonia Toubaline (*), Herv´e Borrion, Tanya Le Sage, Timothy Mitchener-Nissen. 2. Enhancing Portfolio Risk Analysis - Balancing the Past, the Future, the Objective, the Subjective [16:55 - 17:20] by Chris Myers (*). 3. Adversarial Risk Analysis on Transport Infrastructures [17:20 - 17:45] by Javier Cano (*), David R´ıos Insua, Alessandra Tedeschi, Ugur Turhan. 4. Sensitivity and Efficiency of Multiple Criteria Alternatives of the Model ’Revenue Risk’ [17:45 - 18:10] by Rafael Sarkisyan (*), Olga Eskova. Session MO5E (PhD Awards) Room E (James M. Buchanan) - 18:10 MCDM Doctoral Dissertation Award Chaired by Francisco Ruiz 1. Advanced Metaheuristics for Multi-objective Optimization: Design, Analysis and Application [18:10 - 18:15] by Juan J. Durillo (*). 2. New Directions in Robustness Analysis and Preference Modeling in Multiple Criteria Decision Aiding [18:15 - 18:20] by Milosz Kadzinski (*). 3. Hybrid Evolutionary Multi-Objective Optimization for Enhanced Convergence and Diversity [18:20 - 18:25] by Karthik Sindhya (*). 11 Session MO5H (poster) Posters Room H (Lobby) - 18:10 1. Induction of Belief Rules and Belief Decision Trees from Uncertain Data by Khalil AbuDahab (*), Yu-Wang Chen, Ling Xu. 2. Systemic Decision Making: A New Holistic Approach in Ahp-Multiactor Decision Making by Alfredo Altuzarra (*), Pilar Gargallo, Jos´e Mar´ıa Moreno-Jim´enez, Manuel Salvador. 3. Impact of Food Price Volatility over Irrigation Agriculture using a Revealed Preference Model by Carlos Gutierrez-Martin (*), Julio Berbel, Carlos Mario G´omez. 4. Use of Fuzzy ANP-VIKOR Technique in Multi-Criteria Decision-Making of Satisfaction Evaluation for Business-to-Customer (B2C) Services Improvement by Yuching Chern (*), Gwo-Hshiung Tzeng. 5. A Comparative Study of Four Different Multiple Criteria Analysis Methods through the Evaluation of Optimal Locations for C&D Waste Management Facilities by Elena Dosal Vi˜ nas (*), Ana Andr´es Pay´an, Javier Viguri Fuente. 6. Multi-Objective Optimization Problems in Statistical Machine Translation by Kevin Duh (*), Baskaran Sankaran, Anoop Sarkar. 7. Advanced Metaheuristics for Multi-objective Optimization: Design, Analysis and Application by Juan J. Durillo (*). 8. Routing System Optimization – Case Study by Ever Angel Fuentes Rojas (*), Jeyson Andres Mart´ınez Gamboa. 9. Decision Making using Immune Algorithm for Increasing of Level Crossing Capacity by Mikhail Gorobetz (*). 10. A Multi Criteria Decision Analysis-Based Methodology for the Risk Assessment of Flood Hazards at the Regional Scale by Panagiotis Isigonis (*), Andrea Critto, Valentina Gallina, Antonio Marcomini, Elena Semenzin, Silvia Torresan, Alex Zabeo. 11. New Directions in Robustness Analysis and Preference Modeling in Multiple Criteria Decision Aiding by Milosz Kadzinski (*). 12. Precise Consistency Consensus Matrix in an AHP-Group Decision Making Local Context by Jos´e Mar´ıa Moreno-Jim´enez (*), Juan Aguar´on, Mar´ıa Teresa Escobar. 13. Cautious Solutions in Multi-Scenario Bargaining. An Application to Union-Firm Negotiation. by Victoriana Rubiales (*), Amparo Mar´ıa M´armol, Luisa Monroy. 12 14. A Bayesian Negotiation Model for Life Testing by Mar´ıa Jes´ us Rufo (*), Jacinto Mart´ın, Carlos Javier P´erez. 15. DR-MOEA/D: A new Double Reference Point Approach of MOEA/D Based in Aspiration and Reservation Levels by Rub´en Saborido-Infantes (*), Enrique Alba, Mariano Luque. 16. Hybrid Evolutionary Multi-Objective Optimization for Enhanced Convergence and Diversity by Karthik Sindhya (*). 17. Distribution of Federal Funds to Mexican Subnational Governments: Identifying Tools to Enhance MCDA Process in the Public Security Sector by E. Ivonne Vergara-Maldonado (*), Gabriel S´anchez-Guerrero. 13 Tuesday - 18th June Session TU1A (plenary) Plenary 2 Chaired by Jyrki Wallenius Room A (Paraninfo) - 09:00 1. Goal Programming: An Overview of Current Developments and an Application to Offshore Wind Farm Modelling [09:00 - 10:00] by Dylan Jones (*). Session TU2B (contributed) Group Decision Making Chaired by Carlos Romero Room B (Milton Friedman) - 10:20 1. Relational Multi-Attribute Models in DEX Methodology [10:20 - 10:45] by Trdin Nejc (*), Marko Bohanec. 2. What Matters to Stakeholders and Citizens in the Evaluation of Local Policies? Using MCDA to Understand Different Perspectives in Town Renewal [10:45 - 11:10] by Ricardo da Silva Vieira (*), Paula Antunes. 3. Deriving Priority Weights from Pairwise Comparison Matrices under Different Rationality Scenarios [11:10 - 11:35] by Carlos Romero (*), Jacinto Gonz´alez-Pach´on. Session TU2C (contributed) Behavioural Issues (I) Chaired by Gilberto Montibeller Room C (Paul A. Samuelson) - 10:20 1. The Relative Importance of the Criteria and the Ratio of the Directional Derivatives [10:20 - 10:45] by Rafael Sarkisyan (*), Olga Eskova. 2. An Evaluation of the Situation Judgment Test as a Multidimensional Measure of Decision Making [10:45 - 11:10] by Lauren Reinerman-Jones (*), Avonie Parchment, Grace Teo. 3. Strategic Multi-Criteria Decision Analysis [11:10 - 11:35] by Gilberto Montibeller (*). 14 Session TU2D (invited) Room D (Franco Modigliani) - 10:20 Decision/Optimization Models based on Soft Computing (I) Chaired by Carlos Cruz Corona 1. Fuzzy Meta-Goals in Goal Programming [10:20 - 10:45] by Mariano Jim´enez L´opez (*), Mar Arenas-Parra, Amelia Bilbao. 2. Multi-Objective Restricted Dynamic Vehicle Routing Problem with Time Windows [10:45 - 11:10] by J´esica De Armas (*), Bel´en Meli´an, Jos´e A. Moreno P´erez. 3. Fuzzy Linguistic Multi-Criteria Morphological Analysis applied to Scenario Planning [11:10 - 11:35] by Carlos Cruz Corona (*), Mar´ıa T. Lamata, Antonio D. Masegosa, Pablo Villacorta. Session TU2E (invited) Robust Ordinal Regression (I) Chaired by Salvatore Greco Room E (James M. Buchanan) - 10:20 1. Information Levels in Additive Group Decision Models under Incomplete Information: Bridging the Cardinal-Ordinal Gap [10:20 - 10:45] by Luis Dias (*), Paula Sarabando, Rudolf Vetschera. 2. Non Additive Robust Ordinal Regression for Urban and Territorial Planning: an Application for Siting an Urban Waste Landfill [10:45 - 11:10] by Marta Bottero (*), Silvia Angilella, Salvatore Corrente, Valentina Ferretti, Salvatore Greco, Isabella Lami. 3. Entropy Measures to Control Robustness in Ordinal Regression Models [11:10 11:35] by Salvatore Greco (*), Yannis Siskos, Roman Slowinski. Session TU2F (contributed) EMO: Economic Aplications Chaired by Theodor Stewart Room F (Lawrence R. Klein) - 10:20 1. Dealing with Migratory Flows under Uncertainty: a Multi-objective Optimization Approach [10:20 - 10:45] by Pilar Campoy-Mu˜ noz (*), Carlos R. Garc´ıa-Alonso, Melania Salazar-Ordo˜ nez. 2. Robust Selection of Environmental Quality Metrics based on a Multi-Criteria Pareto Data Mining Process [10:45 - 11:10] by Angel Udias (*), Javier Cano, Mauricio Chiazzaro, Lorenzo Galbiati, Andr´es Redchuk. 3. Multiobjective Land-Use Planning Based on Proximity Measures [11:10 - 11:35] by Theodor Stewart (*). 15 Session TU2G (invited) Room G (Robert M. Solow) - 10:20 Multi-Criteria Optimization for the Sustainable Design and Operation of Dynamic Processes (I) Chaired by Filip Logist 1. Multi-Objective Optimization of Integrated Design and Control of a Paper Mill [10:20 - 10:45] by Ingrida Steponavice (*), Kaisa Miettinen. 2. The Treatment of Actuator Failures in Mechatronic Systems by Multi-objective Optimization Methods [10:45 - 11:10] by Christian Horenkamp (*), Michael Dellnitz, Sina Ober-Bl¨obaum, Robert Timmermann. 3. Multi-Objective Optimisation for the Modeling and Control of Uncertain Dynamic (Bio)Chemical Processes [11:10 - 11:35] by Filip Logist (*), Thomas Coenen, Ioanna Stamati, Dries Telen, Mattia Vallerio, Jan Van Impe. Session TU3B (invited) AHP/ANP for Sustainability Chaired by Birsen Karpak Room B (Milton Friedman) - 11:55 1. A Strategic Information System for Sustainable Supply Chain [11:55 - 12:20] by Steve Taraszewski (*), Birsen Karpak. 2. Sustainability Assessment and Spatial ANP: a Methodological Proposal for Studying the New Megacity Region Turin-Milan [12:20 - 12:45] by Valentina Ferretti (*), Marta Bottero, Giulio Mondini. 3. Multi-Criteria Decision Making Approach for Resource Efficient Bread Supply Chains [12:45 - 13:10] by Aleksander Banasik (*), G.D.H. (Frits) Claassen, Argyris Kanellopoulos, Jack van der Vorst. 4. Prioritizing Third Party Logistics Providers: An Analytical Network Process Approach [13:10 - 13:35] by Birsen Karpak (*), Ozden Bayazit. 16 Session TU3C (contributed) MCDM Applications (I) Chaired by Juan Gayt´an-Iniestra Room C (Paul A. Samuelson) - 11:55 1. On Application of Multi-Criteria Decision Making with Ordinal Information in Elementary Education [11:55 - 12:20] by Jir´ı Mazurek (*). 2. The Decision Making Model of Service Improvements of Smart Home System based on IOA-NRM Approach [12:20 - 12:45] by Chung-An Huang (*). 3. Measuring Performance of European Countries in the Cultural Domain: an MCDM Approach [12:45 - 13:10] by Seyhan Nisel (*), Rauf Nisel. 4. A Multicriteria Optimization Model for a Humanitarian Logistics Problem: An Integral Approach [13:10 - 13:35] by Juan Gayt´an-Iniestra (*), Rafael Caballero Fern´andez, Christopher Mej´ıa-Argueta, Juli´an Molina Luque, Bego˜ na Vitoriano. Session TU3D (contributed) Room D (Franco Modigliani) - 11:55 Environmental Multi-Criteria Decision Making (II) Chaired by Alfredo G. Hern´andez-D´ıaz 1. Mixed Generation Sensitivity Analysis: An Optimal Scenario Study [11:55 - 12:20] by Reginald Wilson (*). 2. Abstract REMOVED [12:20 - 12:45] 3. Multi-Criteria Decision Support in Paper Production Industry [12:45 - 13:10] by G.D.H. (Frits) Claassen (*), Johanna Gerdessen. 4. A Multi-objective Waste Collection Problem [13:10 - 13:35] by Alfredo G. Hern´andez-D´ıaz (*), Abraham Duarte, Francisco Gort´azar, Miguel ´ Angel Hinojosa, Ana Dolores L´opez-S´anchez. 17 Session TU3E (invited) Room E (James M. Buchanan) - 11:55 New Methods for Multi-Objective Optimization (II) Chaired by Mariano Luque 1. New Developments of Interactive Multiobjective Optimization using DominanceBased Rough Set Approach [11:55 - 12:20] by Benedetto Matarazzo (*), Salvatore Greco, Roman Slowinski. 2. Non-Linear Multi-Objective Optimization via Parallel Algorithms: Solving a Planar Facility Location and Design Problem [12:20 - 12:45] by Juana L´opez-Redondo (*), Jos´e Fern´andez-Hern´andez, Ar´anzazu Gila-Arrondo, Pilar Mart´ınez-Ortigosa. 3. Reference Point Methods and Approximation of Pareto Sets [12:45 - 13:10] by Kai-Simon Goetzmann (*), Christina B¨ using, Jannik Matuschke, Sebastian Stiller. 4. E-NAUTILUS: A Surrogate based Interactive Multi-objective Optimisation Method without Trading Off [13:10 - 13:35] by Karthik Sindhya (*), Mariano Luque, Kaisa Miettinen, Ana Bel´en Ruiz, Francisco Ruiz. Session TU3F (invited) Room F (Lawrence R. Klein) - 11:55 Evolutionary Multiobjective Optimization (EMO-II) Chaired by Dimo Brockhoff 1. Using Choquet Integral as Preference Model in Interactive Evolutionary Multiobjective Optimization [11:55 - 12:20] by Juergen Branke (*), Salvatore Corrente, Salvatore Greco, Roman Slowinski. 2. A New Evolutionary Algorithm for the Bi-objective Ring Star Problem [12:20 12:45] by Carmen Gal´e (*), Herminia I. Calvete, Jos´e A. Iranzo. 3. GP-DEMO: Differential Evolution for Multiobjective Optimization Based on Gaussian Process Models [12:45 - 13:10] by Miha Mlakar (*), Bogdan Filipic, Dejan Petelin, Tea Tuˇsar. 4. An Improved MOPSO Algorithm for Multiobjective Bilevel Linear Problems [13:10 - 13:35] by Maria Jo˜ao Alves (*). 18 Session TU3G (contributed) Room G (Robert M. Solow) - 11:55 Interactive Multi-Objective Optimization (II) Chaired by Carlos Henggeler Antunes 1. A Comparison of Multi-Objective Approaches to Solve a Parallel Machine Scheduling Problem [11:55 - 12:20] by Yeliz Buruk (*), Gulcan Gocuklu. 2. Decision Support for Multi-Attribute Multi-Item Reverse Auctions [12:20 - 12:45] by G¨ ulsah Karakaya (*), Murat Koksalan. 3. Incorporating Stakeholders’ Preferences into Wastewater Infrastructure Planning in Switzerland [12:45 - 13:10] by Jun Zheng (*), Judit Lienert. 4. An Interactive Tool for Multi-Objective Linear Programming [13:10 - 13:35] by Carlos Henggeler Antunes (*), Maria Jo˜ao Alves. Session TU4B (contributed) AHP/ANP & DEMATEL Chaired by Gwo-Hshiung Tzeng Room B (Milton Friedman) - 15:05 1. The Optimal Resort Development Investment Projects with a Replacement Cost Perspective [15:05 - 15:30] by Chiao-Chen Chang (*), T. Tyrone Lin. 2. A Hybrid Dynamic MCDM Model-Tourism Competitiveness Improvement Strategy [15:30 - 15:55] by Gwo-Hshiung Tzeng (*), Ming-Tsang Lu, Hsin-Chuan Peng, Kua-Hsin Peng. 3. Combined Dominance-based Rough Set Approach and Fuzzy DEMATEL for Analyzing the Adoption Intention of Tourist Guide Mobile Applications [15:55 - 16:20] by Yang-Chieh Chin (*). 4. Enhancing Smart Mobile Devices by the Fuzzy DEMATEL based Network Process [16:20 - 16:45] by Gwo-Hshiung Tzeng (*), Chi-Yo Huang, Yu-Sheng Kao. 19 Session TU4C (contributed) Room C (Paul A. Samuelson) - 15:05 Environmental Multi-Criteria Decision Making (III) Chaired by Valentina Ferretti 1. Comparison of Two Methods for Developing a Multicriteria Evaluation System to Assess Animal Welfare [15:05 - 15:30] by Paula Mart´ın Fern´andez (*), Carlos Buxad´e, Joachim Krieter. 2. A Multi-Criteria Aiding Tool for the Assessment of French Statutory Urban Planning based on ELECTRE TRI. A Case Study of the Local Plan of Toulouse, France [15:30 - 15:55] by Aur´elie Prevost (*), Jean Bandet, Nathalie Molines. 3. Incorporation and Communication of Uncertainty in Decision Support for Environmental Management: A Case Study on Water Quality Improvement [15:55 - 16:20] by Nele Schuwirth (*), Peter Reichert, Christian Stamm. 4. Best Practices for Spatial Multicriteria Evaluation: a Critical Review [16:20 - 16:45] by Valentina Ferretti (*), Gilberto Montibeller. Session TU4D (contributed) Room D (Franco Modigliani) - 15:05 Multi-Objective Optimization Methods (I) Chaired by Jerzy Michnik 1. A Stochastic Choquet Integral Preference Model: SMAA-Choquet [15:05 - 15:30] by Salvatore Corrente (*), Silvia Angilella, Salvatore Greco. 2. Multiobjective Variational Problems Involving Generalized Higher Order Functions [15:30 - 15:55] by Kalpana Shukla (*). 3. Multiobjective Optimization with Combined Global and Local Metamodeling [15:55 - 16:20] by Shapour Azarm (*), Weiwei Hu, Khaled Saleh. 4. Multi-project scheduling as a MCDM problem [16:20 - 16:45] by Bogumila Krzeszowska (*), Tadeusz Trzaskalik. 20 Session TU4E (contributed) Room E (James M. Buchanan) - 15:05 Multi-Attribute Utility Theory (II) Chaired by Rakesh Sarin 1. Multiattribute Utility Analysis using Strict Preference Relations Elicited from a Decision Maker [15:05 - 15:30] by Nishizaki Ichiro (*), Ohmi Masakazu, Hayashida Tomohiro. 2. Using MACBETH with the Choquet Integral to Model Interdependencies between Indicators in the Context of Risk Management [15:30 - 15:55] by Diana F. Lopes (*), Carlos Bana e Costa, M´onica Duarte Oliveira. 3. Determinants of Experienced Utility: Laws and Implications [15:55 - 16:20] by Rakesh Sarin (*), Manel Baucells. Session TU4F (contributed) Room F (Lawrence R. Klein) - 15:05 EMO: Expensive Evaluation Functions Chaired by Julia Handl 1. Optimizing Multi-objective Mixed Integer Programs using Evolutionary Algorithms [15:05 - 15:30] by Thomas Stidsen (*). 2. Automating Mechanism Design using Evolutionary Multi-Criteria Optimisation [15:30 - 15:55] by Arjun Chandra (*), Richard Allmendinger, Peter Lewis. 3. A Neural Network based Hybrid Evolutionary Multiobjective Optimization Algorithm for Computationally Demanding Problems [15:55 - 16:20] by Jussi Hakanen (*), Tommi Kokko, Karthik Sindhya. 4. Redundant Phenotype-Objective Space Mappings in Multi-Criterion Data-Clustering [16:20 - 16:45] by Julia Handl (*), Agah John, Joshua Knowles. 21 Session TU4G (invited) Room G (Robert M. Solow) - 15:05 Socially Responsible Investment Decision Making (II) Chaired by Blanca P´erez-Gladish 1. Practical Measurement of Environmental and Social Achievement Levels for Mutual Funds: A Synthetic Indicator with its Software Application [15:05 - 15:30] by David Pla Santamaria (*), Milagros Bravo Sell´es, Ana M. Garc´ıa Bernabeu, Blanca P´erez-Gladish. 2. Assessing the Efficiency of Equity Mutual Funds. A Fuzzy DEA Approach [15:30 15:55] by Ismael Baeza-Sampere (*), Vicente Coll-Serrano, Bouchra M’Zali, Paz M´endez Rodr´ıguez. 3. Portfolio Multiple Criteria Analysis, Ranking and Optimal Selection [15:55 - 16:20] by Cristinca Fulga (*). 4. A Compromise Programming Model for Mutual Funds’ Socially Responsible Portfolio Selection [16:20 - 16:45] by Blanca P´erez-Gladish (*), Rosario Balaguer, Ana M. Garc´ıa Bernabeu, Paz M´endez Rodr´ıguez. Session TU5B (contributed) AHP/ANP in Practice (I) Chaired by Michael Bruhn Barfod Room B (Milton Friedman) - 17:05 1. A Framework for Sustainable for Third Party Logistic Supplier Identification [17:05 - 17:30] by Birsen Karpak (*). 2. Assessment of Renewable Energies based on Fuzzy and Qualitative Multi-Criteria Decision Making: a Comparison of two Wethodologies [17:30 - 17:55] by Arayeh Afsordegan (*), N´ uria Agell, L´azaro V. Cremades, Monica Sanchez, Siamak Zahedi. 3. Abstract REMOVED [17:55 - 18:20] 4. Combining MCDA and Risk Analysis: Dealing with Scaling Issues in the Multiplicative AHP [18:20 - 18:45] by Michael Bruhn Barfod (*), Kim Bang Salling, Rob Van Den Honert. 22 Session TU5C (contributed) Room C (Paul A. Samuelson) - 17:05 Environmental Multi-Criteria Decision Making (IV) Chaired by Valerie Belton 1. How to Support Environmental Management by MCDM Techniques? [17:05 - 17:30] by Peter Reichert (*), Simone Langhans, Judit Lienert, Nele Schuwirth. 2. Optimizing Economic and Environmental Performances of Solar Power and Electric Vehicles: A MOMILP Application [17:30 - 17:55] by Ellen De Schepper (*), Xavier Gandibleux, Sebastien Lizin, Benjamin Martin, Steven Van Passel, Thomas Vincent. 3. A MAUT Approach for the Management and Remediation of a Site Contaminated by Uranium Processing Activities [17:55 - 18:20] by Miguel C. Mart´ın (*), Antonio Jim´enez, Alfonso Mateos, Danyl P´erez-S´anchez. 4. Climate Change - a Challenge for MCDA? [18:20 - 18:45] by Valerie Belton (*). Session TU5D (invited) Room D (Franco Modigliani) - 17:05 Decision Support for Supply Chain Risk Management Chaired by Tina Comes 1. A Hybrid Method for Supplier Order Quantity Allocation [17:05 - 17:30] by R. J. Kuo (*), C. M. Bai, T. L. Hu. 2. Price-setting Under Risk Criteria for Virtual Products and Services [17:30 - 17:55] by Mordecai Henig (*), Tal Avinadav, Tatyana Chernonog. 3. Minmax Robustness for Multi-Objective Optimization Problems [17:55 - 18:20] by Jonas Ide (*), Matthias Ehrgott, Anita Sch¨obel. 4. Scenario-based Multi-Criteria Decision Support for Robust Humanitarian Relief Supply Chains [18:20 - 18:45] by Tina Comes (*), Frank Sch¨atter. 23 Session TU5E (contributed) MCDM Applications (II) Chaired by Ahmet Kandakoglu Room E (James M. Buchanan) - 17:05 1. Multiobjective Optimization of High-Performance Reinforced Concrete I-Beams by Simulated Annealing [17:05 - 17:30] by V´ıctor Yepes (*), Juli´an Alcal´a, Tatiana Garc´ıa-Segura. 2. Multi-Criteria IMRT Optimisation based on the Generalised EUD Model [17:30 17:55] by Guillermo Cabrera G. (*), Matthias Ehrgott, Andrew Mason. 3. Comparing the Scalarization Methods for Bi-Objective Assembly Line Scheduling Problem [17:55 - 18:20] by Emine Akyol (*), Tugba Sara¸c. 4. A Performance Monitoring System based on a Combined MCDM Methodology [18:20 - 18:45] by Ahmet Kandakoglu (*), Makbule Kandakoglu. Session TU5F (contributed) Room F (Lawrence R. Klein) - 17:05 Multi-Objective Optimization Methods (II) Chaired by Serpil Sayin 1. Improved Constraint Handling Technique for Multi-Objective Optimization [17:05 - 17:30] by Vaibhav Maheshwari (*), Gade Pandu Rangaiah, Shivom Sharma. 2. Computing Nadir Point for Multiobjective Discrete Optimization Problems [17:30 17:55] by Gokhan Kirlik (*), Serpil Sayin. 3. Solving Linear Fractional Programming Problems Using Multiobjective Optimization Techniques [17:55 - 18:20] by Maryam Hasannasab (*), Matthias Ehrgott. 4. Identifying Supported and Extreme Supported Nondominated Solutions and an Empirical Analysis of their Representative Quality [18:20 - 18:45] by Serpil Sayin (*). 24 Session TU5G (invited) Room G (Robert M. Solow) - 17:05 Multi-Criteria Shortest Path Problems: Theory and Applications (I) Chaired by Lorenzo Mandow 1. Engineering Multiobjective Shortest Path Heuristics [17:05 - 17:30] by Christos Zaroliagis (*). 2. Multicriteria Evacuation Plan for Natural Disasters [17:30 - 17:55] by Ismaila Abderhamane Ndiaye (*), Emmanuel Neron. 3. Engineering Parallel Bi-Criteria Shortest Path Search [17:55 - 18:20] by Stephan Erb (*). 4. Searching Graphs with Lexicographic Goal References [18:20 - 18:45] by Lorenzo Mandow (*), Jos´e L. P´erez-de-la-Cruz, Francisco Javier Pulido. 25 Wednesday - 19th June Session WE1B (contributed) AHP/ANP in Practice (II) Chaired by Valerio Salomon Room B (Milton Friedman) - 09:00 1. An Integrated AHP&VIKOR Method for Hospital Site Selection [09:00 - 09:25] by Hafize Yilmaz (*), Sait G¨ ul, Ilker Topcu. 2. Assessment of the Relationship between Planned Obsolescence and Product EcoDesign based on MCDA [09:25 - 09:50] by Pedro Nicol´as Casas P´aez (*), Felix Antonio Cortes Aldana. R The AHP Project on the Cloud [09:50 - 10:15] 3. Elephas: ´ by Fernando P´erez-Rodr´ıguez (*), Angeles Camino-Saco, Esteban G´omez-Garc´ıa, Alejandro Mosquera, Alberto Rojo-Alboreca. 4. Standardization of Classroom Furniture in a Higher Education Institution [10:15 10:40] by Valerio Salomon (*), Pedro Alonso, Fernando Marins. Session WE1C (contributed) Data Envelopment Analysis (I) Chaired by Pekka Korhonen Room C (Paul A. Samuelson) - 09:00 1. Quality of the Information Provided by Fuzzy Data in the Efficiency Evaluation of the Spanish Textile Industry [09:00 - 09:25] by Ismael Baeza-Sampere (*), Vicente Coll-Serrano. 2. Production Points with the Same Reference Hyperplane in DEA: Applications in Sensitivity Analysis and Returns to Scale Classification [09:25 - 09:50] by Akram Dehnokhalaji (*), Nasim Nasrabadi, Majid Soleimani-Damaneh. 3. An Algorithm to Determine Efficiency Scores in Large DEA Data Sets [09:50 - 10:15] by Gazi Bilal Yildiz (*), Banu Soylu. 4. On a Linear Transformation of Variables in Data Envelopment Analysis [10:15 10:40] by Pekka Korhonen (*), Abolfazl Keshvari. 26 Session WE1D (invited) Room D (Franco Modigliani) - 09:00 Multi-Criteria Decision Making in the Energy Sector Chaired by Valentin Bertsch 1. Analytical Approximation of the Influence of Capacity on the Economic and Ecological Performance of Bioenergy Plants [09:00 - 09:25] by Lars-Peter Lauven (*). 2. Multi-Criteria Decision Support for Designing Critical Infrastructure Stress Tests [09:25 - 09:50] by Tina Comes (*), Valentin Bertsch. 3. Simulation-Based Innovization for Improving Productivity and Energy Efficiency of Manufacturing Systems [09:50 - 10:15] by Amos Ng (*), Kalyanmoy Deb, Catarina Dudas, Ainhoa Goienetxea, Matias Urenda Moris. 4. On the Need for Multi-Criteria Modelling in Power Systems Analysis and Planning [10:15 - 10:40] by Valentin Bertsch (*), Wolf Fichtner. Session WE1E (invited) Room E (James M. Buchanan) - 09:00 Promoting Sustainability with MCDM Chaired by Luiz F. Autran M. Gomes 1. Green Economy in the State of Rio de Janeiro - A Multicriteria Evaluation [09:00 09:25] by Rogerio Valle (*), Jo˜ao Cl´ımaco. 2. Sustainable Supply Chains: Contributions from MCDM [09:25 - 09:50] by Birsen Karpak (*). 3. MCDA-GIS Integration for the Evaluation of the Invasive Alien Species. The Case of the Eastern Grey Squirrel in Umbria (Central Italy) [09:50 - 10:15] by Lucia Rocchi (*), Gianluca Massei, Daniele Paoloni. 4. Promoting Sustainability in Mining: Multicriteria Assignment of Resources in Vale [10:15 - 10:40] by Luiz F. Autran M. Gomes (*), Marcelo G. C. Macedo, Roberto Camanho, Thomaz M. Camanho, Luiz Geraldo Biagioni Martins. 27 Session WE1F (invited) Room F (Lawrence R. Klein) - 09:00 Decision/Optimization Models based on Soft Computing (II) Chaired by Carlos Cruz Corona 1. A Decision Support System for Fuzzy Portfolio Selection [09:00 - 09:25] by Carlos Ivorra (*), Clara Calvo, Vicente Liern. 2. Explicit Equations for Non-Dominated Frontier in a Portfolio Selection Problem based on Soft Computing [09:25 - 09:50] by Alberto Carlos Salda˜ na (*), Jos´e Manuel Cadenas, Maria del Carmen Garrido, Raquel Martinez. 3. Fuzzy Multi-Criteria Decision Models for the Portfolio Selection Problem [09:50 10:15] by Enriqueta Vercher (*), Jos´e D. Berm´ udez. 4. Fuzzy Decentralized Production and Distribution Planning in an Uncertain Environment [10:15 - 10:40] by Manuel Diaz-Madro˜ nero (*), Josefa Mula, David Peidro. Session WE1G (contributed) Room G (Robert M. Solow) - 09:00 Multi-Objective Linear Programming Chaired by Petr Fiala 1. About an Implementation of a 3-Objective Linear Programming Solver [09:00 09:25] by Sebastian Schenker (*). 2. Robust Weighted Solutions for Multi-Objective Linear Programming [09:25 - 09:50] by Wlodzimierz Ogryczak (*). 3. A Semidefinite Programming Approach for Solving Multiobjective Linear Programming [09:50 - 10:15] by Victor Blanco (*), Safae El-Haj Ben-Ali, Justo Puerto. 4. Dynamic Multi-Objective De Novo Linear Programming [10:15 - 10:40] by Petr Fiala (*). 28 Session WE2B (contributed) Outranking Methods Chaired by Roman Slowinski Room B (Milton Friedman) - 11:00 1. Rank Reversal in the PROMETHEE Methods: a Summary of Recent Investigations [11:00 - 11:25] by Yves De Smet (*). 2. The SMAA-PROMETHEE Methods [11:25 - 11:50] by Salvatore Corrente (*), Jose Figueira, Salvatore Greco. 3. Measuring Attractiveness of High School Programs using Composite Index of Attractiveness and a Multi-level Outranking Framework: Evidence from PISA Survey in Serbia [11:50 - 12:15] by Mladen Stamenkovic (*), Ivan Anic, Marko Backovic, Zoran Popovic. 4. Solving Ranking Problems with ELECTRE-III in Case of Hierarchical Family of Criteria [12:15 - 12:40] by Luis Del Vasto (*), Roman Slowinski, Aida Valls, Piotr Zielniewicz. Session WE2C (contributed) Applications of Fuzzy Sets Chaired by Ling Xu Room C (Paul A. Samuelson) - 11:00 1. A Fuzzy Numbers Regression for Inducing OECD Wellbeing to WCY Competitiveness 2012 [11:00 - 11:25] by Hsin-Hui Wen (*), Yu-Chien Ko, Gwo-Hshiung Tzeng. 2. A Multi-Criteria Decision Analysis Methodology based on Multi-Attribute Value Theory and Fuzzy Logic for Quantitatively Scoring the Reliability of Ecotoxicological Data [11:25 - 11:50] by Panagiotis Isigonis (*), Philippe Ciffroy, Andrea Critto, Antonio Marcomini, Elena Semenzin, Alex Zabeo. 3. The Evaluation of Competitiveness of Visegrad Four NUTS 2 Regions by the Hybrid Eigenvalue-Fuzzy Cognitive Map Approach [11:50 - 12:15] by Zuzana Kiszova (*), Jir´ı Mazurek, Jan Nevima. 4. An Extension of the Fuzzy DEMATEL Method for Group Decision Making [12:15 - 12:40] by Tarifa Almulhim (*), Ludmil Mikhailov, Ling Xu. 29 Session WE2D (invited) Room D (Franco Modigliani) - 11:00 Applications & Developments of Multicriteria Decision Models Chaired by Marcelo Alencar 1. A MCDA Model to Support Public Safety Management Policy [11:00 - 11:25] by Caroline Mota (*), Andr´e Gurgel. 2. Plaster Waste Destination Problem in the Buildings Sites: An Application of the VFT Methodology [11:25 - 11:50] by Luciana Hazin (*), Marcelo Alencar, Caroline Mota. 3. A Multi-Objective Genetic Algorithm for Integrated Scheduling of Production and Maintenance Activities for a Single Machine [11:50 - 12:15] by Rodrigo Ferreira (*), Adiel Almeida-Filho, Cristiano Cavalcante. 4. Natural Gas Pipelines: A Decision Support System to Evaluate Multidimensional Risk [12:15 - 12:40] by Marcelo Alencar (*), Adiel Almeida, Th´arcylla Clemente, Eduardo Krym. Session WE2E (contributed) MCDM Applications (III) Chaired by Enrique Alba Room E (James M. Buchanan) - 11:00 1. Mixed Integer Multi-Objective Optimization for Flight and Maintenance Planning of Mission Aircraft [11:00 - 11:25] by Andreas Gavranis (*), George Kozanidis. 2. Multicriteria Modeling and Optimization of a Market Place of Leads [11:25 - 11:50] by Maamar Manel (*), Vincent Mousseau, Wassila Ouerdane. 3. Optimizing Telecommunications in Vehicular Networks with a Parallel Multiobjective PSO [11:50 - 12:15] by Jamal Toutouh (*), Enrique Alba. 4. Abstract REMOVED [12:15 - 12:40] 30 Session WE2F (invited) Room F (Lawrence R. Klein) - 11:00 Multi-criteria Shortest Path Problems: Theory and Applications (II) Chaired by Lucie Galand 1. Multivalued Heuristics for One-To-One Biobjective Shortest Path Problems [11:00 - 11:25] by Enrique Machuca (*), Lorenzo Mandow. 2. Multi-Criteria Timetable Information - Extending the Pareto Approach [11:25 11:50] by Mathias Schnee (*), Mohammad Hossein Keyhani. 3. Solving Robust Bicriteria Shortest Path Problems [11:50 - 12:15] by Andrea Raith (*), Kenneth Kuhn, Marie Schmidt, Anita Sch¨obel. 4. Bidirectional versus Unidirectional Heuristic Search for Multiojective Optimization in State Space Graphs [12:15 - 12:40] by Lucie Galand (*), Anisse Ismaili, Patrice Perny, Olivier Spanjaard. Session WE2G (invited) Room G (Robert M. Solow) - 11:00 Strategic Decisions with Multiple Criteria Chaired by Amparo Mar´ıa M´armol 1. An Evaluation of Website Upgrade Options. A Case Study Comparison of ANFIS and RIMER [11:00 - 11:25] by Ling Xu (*), Yu-Wang Chen, Andrada Sabin. 2. Location Decision for Firm Expansion: A Bicriterion Approach [11:25 - 11:50] by Blas Pelegr´ın (*), Pascual Fern´andez Hern´andez, Mar´ıa Dolores Garc´ıa P´erez. 3. Estimating the Pareto Front of a Hard Bi-Criterion Competitive Location Problem [11:50 - 12:15] by Algirdas Lancinskas (*), Pascual Fern´andez Hern´andez, Blas Pelegr´ın, Julius Zilinskas. 4. Pareto Equilibria in Cournot Competition Under Uncertainty [12:15 - 12:40] by M. Angeles Caraballo (*), Eva Buitrago, Amparo Mar´ıa M´armol, Luisa Monroy. 31 Thursday - 20th June Session TH1A (plenary) Plenary 3 Chaired by Pekka Korhonen Room A (Paraninfo) - 09:00 1. Public Decisionmaking and Decision Conferencing [09:00 - 10:00] by Carlos Bana e Costa (*). Session TH2B (contributed) Room B (Milton Friedman) - 11:00 Bi-Criteria Optimization: Methodological Aspects Chaired by Kim Allan Andersen 1. A Two Objective Classification Approach based on Conic Functions [11:00 - 11:25] by Gurkan Ozturk (*), Refail Kasimbeyli. 2. Improved Upper Bounds for a Two-Phase Biobjective Shortest Path Algorithm [11:25 - 11:50] by F. Antonio Medrano (*), Richard Church. 3. A Bi-Objective Mixed-Binary Set Covering Problem [11:50 - 12:15] by Banu Soylu (*). 4. The Bi-Criterion Adaptive Stochastic Knapsack Problem [12:15 - 12:40] by Kim Allan Andersen (*), Matthias Ehrgott, Lars Relund Nielsen, Daniele Pretolani. Session TH2C (contributed) Data Envelopment Analysis (II) Chaired by Josef Jablonsky Room C (Paul A. Samuelson) - 11:00 1. Evaluating Clinical Performance of Pediatric Emergency Physicians with a Help of Data Envelopment Analysis Model [11:00 - 11:25] by Wojtek Michalowski (*), Ken Farion, Fiallos Javier, Jonathan Patrick. 2. Ranking Candidates through Variable Convex Sequences of Weights [11:25 - 11:50] by Bonifacio Llamazares (*). 3. Value Efficiency: A General Case [11:50 - 12:15] by Majid Soleimani-Damaneh (*). 4. Re-Measuring the Happy Planet Index using DEA and MCDM Models [12:15 12:40] by Josef Jablonsky (*). 32 Session TH2D (invited) Fuzzy Goal Programming Chaired by Mariano Jim´enez L´opez Room D (Franco Modigliani) - 11:00 1. An Optimization Model for Group Decision Making: Weighting and Ranking with Incomplete Fuzzy Preference Relations [11:00 - 11:25] by Mauricio Ruiz-Tagle (*), Esther Dopazo. 2. Comparing Fuzzy Goal Programming Approaches for Collaborative Supply Chain Master Planning [11:25 - 11:50] by Manuel Diaz-Madro˜ nero (*), Josefa Mula, David Peidro. 3. Solving Fuzzy Mathematical Programming Problems by a Parametric Approach: a Historical View [11:50 - 12:15] by Carlos Cruz Corona (*), Ricardo Coelho Silva, Jos´e Luis Verdegay Galdeano. 4. A Group Decision Making Model Based in Goal Programming with Fuzzy Hierarchies. An Application to Regional Forest Planning [12:15 - 12:40] by Mariano Jim´enez L´opez (*), Mar Arenas-Parra, Amelia Bilbao, Mar´ıa Victoria Rodr´ıguez-Ur´ıa. Session TH2E (invited) Room E (James M. Buchanan) - 11:00 Multi-Objective Optimization of Computationally Expensive Problems Chaired by Jussi Hakanen 1. Distributed Computation of Pareto Points in Multiobjective Programming [11:00 11:25] by Margaret Wiecek (*), Brian Dandurand. 2. Handling Computational Variability in the Evaluation of Objectives: Ideas and Prospects [11:25 - 11:50] by Julia Handl (*), Richard Allmendinger, Joshua Knowles, Anna Lavygina. 3. Properties and Benefit of the Linear Efficient Frontier Approximation in the Objective Space [11:50 - 12:15] by Volker Maag (*), Tabea Grebe, Uwe Nowak. 4. Agent Assisted Interactive Multiobjective Optimization for Computationally Demanding Problems [12:15 - 12:40] by Vesa Ojalehto (*), Kaisa Miettinen, Dmitry Podkopaev. 33 Session TH2F (contributed) Room F (Lawrence R. Klein) - 11:00 Non Linear Multi-Objective Optimization Chaired by Kathrin Klamroth 1. On Continuation Methods for Non-Linear Multi-Objective Optimization [11:00 11:25] by Benjamin Martin (*), Alexandre Goldsztejn, Laurent Granvilliers, Christophe Jermann. 2. Optimization over the Weakly Efficient Set of a Multiobjective Programming Problem Constrained by Quadratic Functions [11:25 - 11:50] by Syuuji Yamada (*), Tamaki Tanaka, Tetsuzo Tanino. 3. Habitat Availability and Spatial Quality as Conflicting Objectives of Multi-use Forest Management [11:50 - 12:15] by Dmitry Podkopaev (*), Artti Juutinen, Adriano Mazziotta, Kaisa Miettinen, Mikko M¨onkk¨onen, Pasi Reunanen, Olli-Pekka Tikkanen. 4. A Multiple Objective View on Robust Optimization and Stochastic Programming with Finite Scenario Sets [12:15 - 12:40] by Kathrin Klamroth (*), Elisabeth K¨obis, Anita Sch¨obel, Christiane Tammer. Session TH2G (invited) Room G (Robert M. Solow) - 11:00 Representation of the Efficient Set with Quality Guarantees Chaired by Luis Paquete 1. An Efficient Box Algorithm for Discrete Tricriteria Optimization Problems [11:00 11:25] by Kerstin D¨achert (*), Kathrin Klamroth. 2. Finite Representation of Non-Dominated Sets in Multi-Objective Linear Programming [11:25 - 11:50] by Matthias Ehrgott (*), Lizhen Shao. 3. Transforming Constraints into Objectives: Experiments with Bidimensional Knapsack Problems [11:50 - 12:15] by Britta Schulze (*), Kathrin Klamroth, Luis Paquete. 4. Finding Representative Subsets in Multiobjective Discrete Optimization [12:15 12:40] by Luis Paquete (*), An´ıbal Ponte, Michael Stiglmayr, Daniel Vaz. 34 Session TH3B (invited) Robust Ordinal Regression (II) Chaired by Roman Slowinski Room B (Milton Friedman) - 14:10 1. Multiple Criteria Hierarchy Process for ELECTRE TRI Methods [14:10 - 14:35] by Salvatore Corrente (*), Salvatore Greco, Roman Slowinski. 2. Stochastic Ordinal Regression for Multiple Criteria Sorting Problems [14:35 - 15:00] by Milosz Kadzinski (*), Tommi Tervonen. 3. Inference of Parsimonious Preference Models in Robust Ordinal Regression [15:00 15:25] by Roman Slowinski (*), Salvatore Greco, Vincent Mousseau. Session TH3C (contributed) Classification Chaired by N´ uria Agell Room C (Paul A. Samuelson) - 14:10 1. Feature Selection for Classification Using Multi-Objective Optimization [14:10 14:35] by Emre Cimen (*), Gurkan Ozturk. 2. Multi-Criteria Preference Disaggregation Analysis for Classification: an Application to Model Colour Preferences [14:35 - 15:00] by Mohammad Ghaderi (*), N´ uria Agell, Monica Sanchez, Francisco Javier Ruiz. 3. Increasing the Transparency of Model Selection in Multi-Criterion Data-Clustering [15:00 - 15:25] by Julia Handl (*), K.Nadia Papamichail. Session TH3D (contributed) Behavioural Issues (II) Chaired by Alexis Tsouki`as Room D (Franco Modigliani) - 14:10 1. A New Dominance Measuring Method for MCDM with Ordinal Information About DM’s Preferences [14:10 - 14:35] by Ernesto Aguayo (*), Antonio Jim´enez, Alfonso Mateos. 2. The Support of Decision Criteria Weights Elicitation based on their Probabilistic Characteristics [14:35 - 15:00] by Jakub Brzostowski (*), Ewa Roszkowska, Tomasz Wachowicz. 3. Primitive and Derived Information in Decision Models [15:00 - 15:25] by Alexis Tsouki`as (*), Alberto Colorni. 35 Session TH3E (invited) Room E (James M. Buchanan) - 14:10 Multi-Criteria Optimization for the Sustainable Design and Operation of Dynamic Processes (II) Chaired by Filip Logist 1. Bi-Criteria Optimisation using Surrogate Modelling for Dynamic Process Design [14:10 - 14:35] by Eric Fraga (*), Joakim Beck. 2. Multi-Objective Framework for Model Based Design of Experiments: Application to Hemodialysis and Type 1 Diabetes Model [14:35 - 15:00] by Vaibhav Maheshwari (*), Gade Pandu Rangaiah, Lakshminarayanan Samavedham. 3. Applying of Fuzzy Logic to Decision-Making Control of the Ship Motion [15:00 15:25] by Jozef Malecki (*). Session TH3F (contributed) EMO for Engineering Design Chaired by Ansgar Dietermann Room F (Lawrence R. Klein) - 14:10 1. A Tabu Search Algorithm for Multi Objective Open Vehicle Routing Problem [14:10 - 14:35] by Erdener Ozcetin (*), Gurkan Ozturk. 2. Efficient Optimisation of a Vehicle Suspension by Hybrid Nature Inspiring Method [14:35 - 15:00] by Darakhshan Jabeen Syeda (*). 3. Structured Development of Automotive Electric/Electronic Architectures Using Evolutionary Optimization [15:00 - 15:25] by Ansgar Dietermann (*), Bernard B¨aker. Session TH3G (invited) Room G (Robert M. Solow) - 14:10 Decision Making for the Use of Natural Resources (I) Chaired by Jos´e Ant´on 1. A Multicriteria Decision Aiding Methodology to Support Public Administration on Sustainable Energy Action Plans [14:10 - 14:35] by Maria Franca Norese (*), Giuliano Dall’O, Chiara Novello. 2. Comparison between Continuous and Discrete Multicriteria Decision Methods applied to Lands Use, Conservation, Management and Planning on Hydro-Basin in Cordoba Province (Argentina) [14:35 - 15:00] by Juan Grau (*), Diego Andina, Jos´e Ant´on, Jos´e Manuel Cisneros, Ana Tarquis. 3. Some Remarks about Multi-Criteria Methods into Planning for Use and Services of Natural Areas [15:00 - 15:25] by Jos´e Ant´on (*), Diego Andina, Jos´e Manuel Cisneros, Juan Grau, Roland Portait, Maria S´anchez, Ana Tarquis. 36 Session TH4B (contributed) Room B (Milton Friedman) - 15:45 AHP/ANP Applications to Development Economics Chaired by Jos´e A. G´omez-Lim´on 1. Decomposing Value Creation when Assessing Investments: a Multi-Criteria Approach [15:45 - 16:10] by Mar´ıa Dolores Guerrero-Baena (*), Jos´e A. G´omez-Lim´on. 2. A System Approach for Green Supplier Selection Based on the ANP Model with BOCR for the Electronics Industry in Taiwan [16:10 - 16:35] by Hsu-Shih Shih (*), Chiau-Ching Chen, Yi-Chun Lin. 3. A Three-Stage Methodology to Analyze the Global Competitiveness of a Sector [16:35 - 17:00] by Ozay Ozaydin (*), Bora Cekyay, Ozgur Kabak, Sule Onsel Ekici, Fusun Ulengin. 4. Analyzing the Provision of Agricultural Public Goods: The Case of Irrigated Olive Groves in Southern Spain [17:00 - 17:25] by Anastasio Villanueva-Rodr´ıguez (*), Manuel Arriaza, Jos´e A. G´omez-Lim´on, Olexandr Nekhay. Session TH4C (contributed) Sustainability Analysis Chaired by Jian-Bo Yang Room C (Paul A. Samuelson) - 15:45 1. New Alternative Aggregation Formulas for the Human Development Index: a Multicriteria Approach [15:45 - 16:10] by Beatriz Rodr´ıguez D´ıaz (*), Mar´ıa Jos´e Angulo-Guerrero, Mariano Luque, Salvador P´erez-Moreno. 2. Design Safer and Greener Road Projects by using a Multi-Objective Optimization Approach [16:10 - 16:35] by Renaud Sarrazin (*), Yves De Smet. 3. Sustainable National Transport Planning – Governance with MCDA? [16:35 - 17:00] by Anders Vestergaard Jensen (*), Steen Leleur, Joe Zietsman. 4. The Derivation of Weights for Sustainability Criteria: A Framework for Corporate Assessment in Sugar Manufacturing [17:00 - 17:25] by Panitas Sureeyatanapas (*), David Bamford, Jian-Bo Yang. 37 Session TH4D (contributed) Room D (Franco Modigliani) - 15:45 Integer Multi-Objective Optimization Chaired by Johanna Gerdessen 1. Web-Based Decision Support System for Solving Multiple Objective Linear Integer Problems (MOLIPs) [15:45 - 16:10] by Leoneed Kirilov (*), Krasimira Genova, Vassil Gouliashki, Boris Staykov. 2. On Branch and Bound Approach to Multi-Objective Global Optimization [16:10 16:35] by Antanas Zilinskas (*), Julius Zilinskas. 3. Multicriteria 0-1 Knapsack Problems with K-min Objectives [16:35 - 17:00] by Aiying Rong (*), Jose Figueira, Kathrin Klamroth. 4. MCDM in Nutritional Epidemiology [17:00 - 17:25] by Johanna Gerdessen (*), Aleksander Banasik, G.D.H. (Frits) Claassen. Session TH4E (contributed) MCDM Applications (IV) Chaired by Jose Roberto Ribas Room E (James M. Buchanan) - 15:45 1. Using Methods of Multi-Criteria Decision Making to Evaluate the Sustainability of Bioenergy Concepts by Integrating Adequate Reference Points [15:45 - 16:10] by Nils Lerche (*), Jutta Geldermann. 2. Web Multicriteria Spatial Decision Support System: Integration of a Web-based ELECTRE TRI in ArcGIS [16:10 - 16:35] by Sandra Silva (*), Lu´ıs Al¸cada-Almeida, Luis Dias. 3. Unsupervised Update of the Preferences on Numerical and Multi-Valued Categorical Criteria from the Analysis of the User Behaviour [16:35 - 17:00] by Aida Valls (*), David Isern, Lucas Marin, Antonio Moreno. 4. An Assessment of Multipurpose Water Resources Using Multicriteria Analysis [17:00 - 17:25] by Jose Roberto Ribas (*), Mariana Pinheiro, Juliana Ribas Severo. 38 Session TH4F (contributed) EMO in Industrial Engineering Chaired by Fouad Ben Abdelaziz Room F (Lawrence R. Klein) - 15:45 1. A Genetic Algorithm for a Multi-Objective Parallel Machine Scheduling Problem with Special Process Constraints [15:45 - 16:10] by Tugba Sara¸c (*), Yakup Emre Aydinbeyli, M¨ uzeyyen Melek Ko¸canli. 2. A Genetic Algorithm Approach for Solving Multi-Objective Factory Layout Problem [16:10 - 16:35] by Phong Ho (*), Truong Huy, Ho Quoc. 3. Multiple Criteria Decision Making in Anti-Collision Scheduling for Level Crossing Tasks [16:35 - 17:00] by Anatoly Levchenkov (*). 4. Multi-Objective Task Allocation using Multi-Agent Coalition Formation [17:00 17:25] by Fouad Ben Abdelaziz (*), Noha Amer. Session TH4G (contributed) Room G (Robert M. Solow) - 15:45 Bi-Criteria Optimization: Applications Chaired by Jacques Teghem 1. A Column Generation based Algorithm for a Bi-Objective Capacity Evaluation of a Railway Infrastructure [15:45 - 16:10] by Evgeny Gurevsky (*), Xavier Gandibleux, Aur´elien Merel. 2. A Bi-Objective Decomposition Method for Solving the Bi-Objective Multi-Commodity Minimum Cost Flow Problem [16:10 - 16:35] by Siamak Moradi (*), Matthias Ehrgott, Andrea Raith. 3. Bi-objective Traffic Assignment with Multiple User Classes: A Time Surplus Approach [16:35 - 17:00] by Olga Perederieieva (*), Matthias Ehrgott, Andrea Raith, Judith Y. T. Wang. 4. A biobjective Parallel Machine Scheduling Problem with Eligibility, Ready Dates and Delivery Times [17:00 - 17:25] by Jacques Teghem (*), Manuel Mateo. 39 Friday - 21st June Session FR1A (plenary) MCDM Society Awardees Session Chaired by Murat K¨oksalan Room A (Paraninfo) - 09:00 Session FR2B (contributed) Room B (Milton Friedman) - 11:50 AHP/ANP: Methodological Issues (II) Chaired by Gang Kou 1. A Bayesian Analysis Method for AHP based on Group Decision Making [11:50 12:15] by Gang Kou (*), Changsheng Lin, Yi Peng. 2. Structural Modeling Approach to Multiple Criteria Problems with Interrelated Components [12:15 - 12:40] by Jerzy Michnik (*). 3. Comparative Analysis and Evaluation of Multi-Criteria Decision Making Methods [12:40 - 13:05] by Gang Kou (*), Daji Ergu, Thomas L. Saaty. Session FR2C (invited) Robust Ordinal Regression (III) Chaired by Vincent Mousseau Room C (Paul A. Samuelson) - 11:50 1. Learning the Parameters of a Multiple Criteria Sorting Method from Large Sets of Assignment Examples [11:50 - 12:15] by Olivier Sobrie (*), Vincent Mousseau, Marc Pirlot. 2. Robust Ordinal Regression in Geographical Information Systems [12:15 - 12:40] by Salvatore Greco (*), Antonio Boggia, Salvatore Corrente, Gianluca Massei, Roman Slowinski. 3. Robust Elicitation of a Qualitative Ranking Model using Inconsistent Data [12:40 13:05] by Wassila Ouerdane (*), Jinyan Liu, Vincent Mousseau. 40 Session FR2D (contributed) Room D (Franco Modigliani) - 11:50 Applications of Multi-Attribute Utility Theory Chaired by Mario Castillo 1. Modeling Consumer Preferences about Vehicles with Multi-Attribute Additive Models: Survey-Based Experiments [11:50 - 12:15] by Luis Dias (*), Gabriela Oliveira, Paula Sarabando. 2. Factors Affecting Marketing Channel Choice Decisions in Citrus Supply Chain: A Conceptual Framework [12:15 - 12:40] by Muhammad Imran Siddique (*), Elena Garnevska, Norman Marr. 3. Use of Multiattribute Utility Function for a Cost-Utility Analysis Kangaroo Mother versus “Traditional” Care for Premature Infants in Bogota, Colombia [12:40 - 13:05] by Mario Castillo (*), Astrid Bernal, Nathalie Charpak, Mar´ıa-Adelaida C´ordoba, John-Jairo R´ıos, Juan-Gabriel Ruiz, Tammy Trujillo. Session FR2E (invited) Room E (James M. Buchanan) - 11:50 Methods with Guaranteed Performance for Multi-Objective Combinatorial Optimization (II) Chaired by Daniel Vanderpooten 1. Two-Phase Method for Multiobjective Mixed 0-1 Linear Programming [11:50 - 12:15] by Thomas Vincent (*), Xavier Gandibleux, Anthony Przybylski. 2. Generation of the Choquet Optimal Set of Multiobjective Combinatorial Optimization Problems [12:15 - 12:40] by Thibaut Lust (*), Antoine Rolland. 3. On Approximate Kernels of Minimal Size for Bicriteria Problems [12:40 - 13:05] by Florian Jamain (*), Cristina Bazgan, Daniel Vanderpooten. Session FR2F (invited) Room F (Lawrence R. Klein) - 11:50 Evolutionary Multiobjective Optimization (EMO-III) Chaired by Dimo Brockhoff 1. Satisfying Multiple Objectives in Wastewater Treatment Network Design Using a Diversity Preserving Evolutionary Algorithm [11:50 - 12:15] by Hana Chmielewski (*), S. Ranji Ranjithan. 2. An Evolutionary Multiobjective Optimization Algorithm for Diverse Solutions [12:15 - 12:40] by Hana Chmielewski (*), Brian E.B. Piper, S. Ranji Ranjithan. 3. Cross Entropy for Combinatorial Optimization Problems with Linear Relaxations [12:40 - 13:05] by Juli´an Molina Luque (*), Rafael Caballero Fern´andez, Alfredo G. Hern´andezD´ıaz, Manuel Laguna. 41 Session FR2G (contributed) Preference Modeling Chaired by Andrzej M.J. Skulimowski Room G (Robert M. Solow) - 11:50 1. Experimental Evaluation of Polynomial and Copula Functions for Qualitative Option Ranking [11:50 - 12:15] by Marko Bohanec (*), Biljana Mileva-Boshkoska. 2. A Preference-Based Facility Location Problem [12:15 - 12:40] by Katarzyna Krupinska (*). 3. Proximity-Based Decision Rules for Multicriteria Model Predictive Control Problems [12:40 - 13:05] by Andrzej M.J. Skulimowski (*). Session FR3B (contributed) Room B (Milton Friedman) - 14:35 Goal Programming/Compromise Programming Chaired by Luis A. Manotas 1. Multicriteria Model for Selecting Renewable Energy Projects [14:35 - 15:00] by Diego F. Manotas (*), Carlos A. Lozano, Carlos Vidal. 2. A Sequential Goal Programming Model for University Timetable Construction [15:00 - 15:25] by Veronika Skocdopolova (*). 3. Multicriteria Model for Locating Distribution Centers [15:25 - 15:50] by Luis A. Manotas (*), Diego F. Manotas , Germ´an Ocampo, Carlos Vidal. 4. A Multi-Objective Route Choice Model for Health-Conscious Cyclists [15:50 - 16:15] by Judith Y. T. Wang (*), Alan K. L. Cheung, Kim N. Dirks, Matthias Ehrgott, Jon Pearce. Session FR3C (contributed) Room C (Paul A. Samuelson) - 14:35 Integer and Combinatorial Multi-Objective Optimization Chaired by Banu Lokman 1. A New Exact Method for Optimizing a Linear Function over an Integer Efficient Set [14:35 - 15:00] by Djamal Chaabane (*). 2. A Fuzzy Model for Project Portfolio Selection [15:00 - 15:25] by F´atima P´erez (*), Rafael Caballero Fern´andez, Trinidad G´omez, Vicente Liern. 3. An Interactive Approach to the Bi-objective Inventory Routing Problem [15:25 15:50] by Sandra Huber (*), Martin Josef Geiger, Marc Sevaux. 4. Finding Nadir Points in Multi-objective Integer Programs [15:50 - 16:15] by Banu Lokman (*), Murat Koksalan. 42 Session FR3D (invited) Room D (Franco Modigliani) - 14:35 Multi-Objective Optimization and Decision Making under Uncertainty Chaired by Juergen Branke 1. Risk-Averse Bi-Objective Optimization via Stochastic Dominance under Transaction Costs [14:35 - 15:00] by Walter Gutjahr (*). 2. Multi-Objective Optimization of Combinatorial Problems with Fuzzy Data [15:00 15:25] by Oumayma Bahri (*), Nahla Ben Amor, Talbi El-Ghazali. 3. Stochastic Approch Versus Optimizition Over Efficient Set [15:25 - 15:50] by Fatma Mebrek (*), Djamal Chaabane. Session FR3E (contributed) Risk MCDM Modeling (II) Chaired by Shuang Liu Room E (James M. Buchanan) - 14:35 1. A Multicriteria Prediction Model for Project Classifications [14:35 - 15:00] by Rueben Laryea (*). 2. Applying the Utadis Method to the Identification of Key Factors of Organisational Commitment: A Case Study in Petrobras [15:00 - 15:25] by Luiz F. Autran M. Gomes (*), Marcos E.L. Gon¸calves, Luis A.D. Rangel. 3. Multi-Criteria Sorting with Category Size Restrictions [15:25 - 15:50] ¨ by Selin Ozpeynirci (*), Murat Koksalan, Vincent Mousseau. 4. Using Structured Decision Making to Facilitate Environmental Risk Management [15:50 - 16:15] by Shuang Liu (*). 43 Session FR3F (contributed) Water Resources Management Chaired by Eun-Sung Chung Room F (Lawrence R. Klein) - 14:35 1. Risk Assessment of Large Scale Hydroelectric Plant Projects using Fuzzy Analytical Hierarchy Process [14:35 - 15:00] by Jose Roberto Ribas (*), Juliana Ribas Severo, Flavio Sohler. 2. Multi-Criteria Decision Analysis for Water Supply Infrastructure Planning under Uncertainty [15:00 - 15:25] by Lisa Scholten (*), Judit Lienert. 3. Rewarding Decisions Beyond the Performance Targets Paradigm [15:25 - 15:50] by Victor Sousa (*), Ana Faustino. 4. Fuzzy VIKOR Approach for Assessing the Vulnerability of the Water Supply to Climate Change and Variability in South Korea [15:50 - 16:15] by Eun-Sung Chung (*), Boram Lee, Kwang-Jae Won. Session FR3G (invited) Room G (Robert M. Solow) - 14:35 Decision Making for the Use of Natural Resources (II) Chaired by Juan Grau 1. Classification of Disasters Effects and Humanitarian Decision-Makers’ Criteria [14:35 - 15:00] by Tinguaro Rodriguez (*), Javier Montero, Bego˜ na Vitoriano. 2. Discussion of the Results Obtained in Several Real Applications using PROMETHE I and II versus Weighted PROMETHE and other Outranking Methods [15:00 - 15:25] by Juan Grau (*), Diego Andina, Jos´e Ant´on, Federico Colombo, Lisandro De los R´ıos, Ana Tarquis. 3. Development of Mathematical Multicriteria Methods to Aid to Elaborate Plans of Rural Sustainable Development in Areas With Aborigines Communities [15:25 15:50] by Jos´e Ant´on (*), Diego Andina, Juan Grau, Jos´e Luis Rubio, Ana Tarquis. 4. Development of an Innovative Environmental Education Program for Urban Solid Waste Management through Multi-Criteria Decision Methods (MCDM) [15:50 16:15] by Juan Grau (*), Diego Andina, Jos´e Ant´on, Rodrigo Martin, Alfredo Soletti, Ana Tarquis. Closing Session Room A (Paraninfo) - 16:15 44 Abstracts 45 FR2B (contributed) A Bayesian Analysis Method for AHP based on Group Decision Making Kou, Gang (*) Lin, Changsheng Peng, Yi UESTC The Analytic Hierarchy Process (AHP) is widely used in Group Decision Making (GDM). The aggregation of individual judgments (AIJ) and the aggregation of individual priorities (AIP) are two of the most widely employed aggregation methods in AHP based on GDM. The overall priority vector can be obtained by many different methods such as weighted geometric mean method (WGMM), goal programming approach (GPA) and decision support model (DSM) etc. Different from these methods, in this paper, we propose a new Bayesian analysis method (BAM) for the AIP, which makes full use of the prior distribution for the parameters and sample information while complying with the Pareto principal. We also develop a multiplicative model with log-normal errors under the hypothesis of the existence of consensus among the decision makers and evaluate the statistical properties of the priority vector obtained by employing the posterior distribution. Besides, a measurement to assess the group consistency is proposed. The proposed BAM provides a flexible framework and efficient estimates for obtaining the priority vector. Three numerical examples are used to illustrate the applications and advantages of the BAM. The results are compared with those produced by the other techniques according to Euclidean distance (ED), Minimum Violation (MV) and Mean absolute deviation (MAD). The results show that the proposed BAM is more flexible, practical and realistic than the other methods. Keywords: Aggregation Of Individual Priorities, Analytic Hierarchy Process, Group Decision Making. 46 MO5H (poster) A Bayesian Negotiation Model for Life Testing Rufo, Mar´ıa Jes´ us (*) Mart´ın, Jacinto P´erez, Carlos Javier University of Extremadura This work presents a Bayesian negotiation model in the context of life testing. A general situation in which a manufacturer offers a product to a consumer is considered. The reliability of the product depends on its lifetime. Furthermore, the consumer can use her/his own lifetime distribution to address the uncertainty about the information provided by the manufacturer with respect to the quality of the product. Hence, the consumer might accept o reject the product in advance. By other hand, the manufacturer would have to convince the consumer of the product reliability. Then, the manufacturer offers the consumer a sample of size n to make life testing. Based on this sample, the consumer must decide to accept or reject the product. In order to solve the problem, a Bayesian sequential model of negotiation is proposed here. Firstly, it is considered that the parametric lifetime distributions belong to a subclass of the exponential family. Thus, a unified framework is carried out. For these distributions, different prior distributions for the manufacturer and the consumer are considered. The utility functions for both, the manufacturer and the consumer, are based on an observable life length. Moreover, the utility function for the manufacturer includes the corresponding costs. A simulation-based approach is implemented to calculate the optimal sample size n that the manufacturer will offer the consumer. Finally, an application is presented and discussed in order to show that the proposed technique is easily applied in practice. Keywords: Bayesian Analysis, Exponential Family, Life Testing, Negotiation. 47 TH4G (contributed) A Bi-Objective Decomposition Method for Solving the Bi-Objective Multi-Commodity Minimum Cost Flow Problem Moradi, Siamak (*) Raith, Andrea The University of Auckland Ehrgott, Matthias Lancaster University We address linear minimum cost flow problems with two objectives and multiple commodities. We present a new method for solving the problem by integrating the Dantzig-Wolfe decomposition method in the iterations of the bi-objective simplex method. The standard bi-objective simplex method is initialized by optimizing the problem with respect to the first objective. The method then iteratively finds entering variables with the maximum ratio of improvement of the second objective over the deterioration of the first objective. The method stops when all the efficient extreme solutions are obtained. Our proposed method has a similar structure where the initial solution can be obtained by solving a single objective multi-commodity flow problem with the existing Dantzig-Wolfe decomposition method. The method then continues iteratively by solving subproblems with only single commodity network constraints and a ratio objective function. The optimal solution of each subproblem which is a variable may be a candidate for entering the basis. Among them we choose the one with the best ratio objective value to ensure the method moves from one extreme point to the next. We also demonstrate that the optimal solution of each subproblem can be found among the efficient solutions of a linear bi-objective problem. Keywords: Bi-Objective Multi-Commodity Flow Problem, Bi-Objective Simplex Method, Decomposition Method. 48 TH2B (contributed) A Bi-Objective Mixed-Binary Set Covering Problem Soylu, Banu (*) Erciyes University In this study, a bi-objective mixed-binary set covering problem is considered and algorithms are developed to find Pareto solutions of this problem. There are two decisions in our case: one is to decide on which sets should be selected and the other is to determine some parameter levels for the selected sets. Although the problem is originally mixed binary nonlinear programming problem, it can be linearized. The Pareto frontier of this problem has some interesting properties. We propose two methodologies: one is based on the decomposition approach and the other is based on the branch-and-bound algorithm. Both methods can also be generalized to find Pareto solutions of mixed integer linear programming (MILP) problems. We present our numerical experiments and findings. Keywords: Branch and Bound, Mixed Integer Programming, Set Covering Problem. 49 TH4G (contributed) A biobjective Parallel Machine Scheduling Problem with Eligibility, Ready Dates and Delivery Times Teghem, Jacques (*) University of Mons Mateo, Manuel Universitat Polit`ecnica de Catalunya The shop has three parallel machines of levels 1, 2 and 3 respectively corresponding to decreasing quality production. The jobs to process on one of the machines have processing times - identical for all the machines -, ready dates and delivery times. These jobs correspond also to one of the three levels meaning that a job of level l is eligible to be assigned only to the machines of levels 1 till l. Nevertheless if a job is assigned to a machine of level k different than 1, a penalty k - 1 is incurred. Two criteria are considerd: the makespan and the total penalty. The aim is to determine or to approximate the set of efficient schedules for the minimization of these two criteria. Keywords: Multi-Objective Optimization, Production Scheduling. 50 MO3E (contributed) A Careful Look at the Importance of Criteria and Weights Wallenius, Jyrki (*) Korhonen, Pekka ¨ orni, Anssi Silvennoinen, Kari O¨ Aalto University We investigate the connection between weights, scales, and the importance of criteria, when a linear value function is assumed to be a suitable representation of a decision maker’s preferences. Our considerations are based on a simple two-criteria experiment, where the participants were asked to indicate which of the criteria was more important, and to pairwise compare a number of alternatives. We use the participants’ pairwise choices to estimate the weights for the criteria in such a way that the linear value function explains the choices to the extent possible. More specifically, we study two research questions: 1) is it possible to find a general scaling principle that makes the rank order of the importance of criteria consistent with the rank order of the magnitudes of the weights, and 2) how good is a simple, direct method of asking the decision maker to “provide” weights for the criteria compared to our estimation procedure. Our results imply that there is reason to question two common beliefs, namely that the values of the weights would reflect the importance of criteria, and that people could reliably “provide” such weights without estimation. Keywords: Behavioral, Criterion Importance, Multi-Attribute Value Theory, Weights. 51 TH4G (contributed) A Column Generation based Algorithm for a Bi-Objective Capacity Evaluation of a Railway Infrastructure Gurevsky, Evgeny (*) Gandibleux, Xavier Merel, Aur´elien University of Nantes A capacity evaluation of a railway infrastructure is one of the important issues of the preliminary planning for the corresponding railway traffic. Given a train schedule and an additional circulation demand, it consists in estimating the maximal number of trains that can pass through the infrastructure while satisfying the possible exploitation constraints and conflicts. In the literature, this task has received the name of saturation problem. In the current context of the European railway market which are going to be open to competing companies, the saturation problem becomes an essential economic challenge. Thus, an efficient management of a railway infrastructure has to be done not only maximizing its capacity but in such a manner as to also minimize the disparity between the companies using this infrastructure. Taking into account this argument, the problem becomes bi-objective. To model this problem, we use the set packing formulation, where the decision variables ask if a train is able to use a railway within a given time-period and the constraints describe all the conflicting situations. In the case of a real-world application, such formulation is of extremely big size due to an exponential number of train assignments. As a consequence, for its exact resolution, a good quality of bound sets is needed. In this work, for the problem considered, we study the linear programming (LP) relaxation that can be viewed as its bound set. To solve the LP relaxation, the column generation (CG) procedure is used. During the presentation, we will mention some difficulties of applying CG for the bi-objective linear programming problems and will show different ways of constructing bound sets based on the information provided. The computational experiments will be carried out on a collection of benchmark data. The results will be compared and discussed. Keywords: Applications of MCDM, Bicriteria Model, Large Scale Optimization. 52 MO5H (poster) A Comparative Study of Four Different Multiple Criteria Analysis Methods through the Evaluation of Optimal Locations for C&D Waste Management Facilities Dosal Vi˜ nas, Elena (*) Andr´es Pay´an, Ana Viguri Fuente, Javier University of Cantabria Specific legislation has been developed in the last years for the improvement of Construction and Demolition Waste (C&DW) management, and therefore, recycling facilities must be located. Selection of a non-appropriate location is a critical aspect which affects long-term profits and can lead into environmental and social impacts as well. For this reason, this decision must be studied thoroughly, taking into account several targets and aspects that usually are opposite. Multiple Criteria analysis (MCA) is usually applied to compare scenarios regarding to contradictory objectives, and a wide range of different methods are defined, with different forms to calculate the total value score for an alternative. In this work, a comparative study of four different MCA methods; Weighted Summation, Regime, Electre II and Evamix; was performed by means its application to select optimal locations for C&DW management facilities in Cantabria, Spain. Results obtained are analyzed in order to assess if the method applied may provide different results with the same data, which may lead to a non-objective identification of the best alternative. Besides, analyses of the sensitivity and uncertainty of the results have been also carried out in order to investigate the robustness of the results obtained. Uncertainty of input data includes uncertainty in data selection and uncertainty in measurement values, whereas uncertainty in the model itself includes model structure and parameters. The main results show that the application of more than one MCA method to verify the results is highly recommended; mainly to locate C&DW recycling facilities due to this decision involves a large investment that must be insured for a long time. Acknowledgments. This work has been supported by R+D project “Establecimiento del Sistema de Indicadores para el Flujo Sostenible de Recursos y Residuos en la Comunidad Aut´onoma de Cantabria”, Government of Cantabria-University of Cantabria, Spain. Keywords: Comparative Analysis, Construction And Demolition Waste, Multi-Criteria Decision Making, Recycling Facilities, Sensitivity Analysis, Uncertainty. 53 TU3G (contributed) A Comparison of Multi-Objective Approaches to Solve a Parallel Machine Scheduling Problem Buruk, Yeliz (*) Eskisehir Osmangazi University Gocuklu, Gulcan Eskisehir Osmangazi University Although there are many problems that are well-known for single objective optimization, the real world problems have more than one conflicting objectives. For such problems, various multi-objective techniques take part in the literature and there is not a known proven optimal solution method for all kinds of problems. In the scope of this study, a multi-objective parallel machine scheduling problem has been considered. WeightedTchebycheff, epsilon-constraint and conic scalarization methods were used to solve the problem by using GAMS v.12.1 software package. A set of solutions was identified to represent the trade-off between different considerations. The computational results for each technique have been compared in terms of efficiency and applicability. Keywords: Mathematical programming, Multi-Objective Optimization, Scalarization, Scheduling. 54 TU4G (invited) A Compromise Programming Model for Mutual Funds’ Socially Responsible Portfolio Selection P´erez-Gladish, Blanca (*) University of Oviedo Garc´ıa Bernabeu, Ana M. Universidad Polit´ecnica de Valencia Balaguer, Rosario Universitat Jaume I M´endez Rodr´ıguez, Paz University of Oviedo Socially Responsible Investing (SRI) corresponds to an investment practice that takes into account not only the usual return-risk criteria, but also other non-financial dimensions, namely in terms of environmental, social and governance concerns. However, while a diverse set of models has been developed to support investment decision-making based on financial criteria like risk and return, models including also socially responsible criteria are rather scarce. In this paper we propose a multicriteria portfolio selection model for mutual funds based on Compromise Programming which takes into account both, financial and SRI criteria. The proposed model is intended to be an individual investment decision making tool for mutual funds’ portfolio selection taking into account the subjective and individual preferences about different financial and non-financial features of an individual investor. Few portfolio selection models include among their decision making criteria non financial ones. In order to do so, the first problem to be solved is the measurement of the degree of social responsibility of a mutual fund. The second one is how to incorporate the former one as a decision criterion in the portfolio selection model and how to elicit from the particular investor his/her personal preferences regarding the different financial and non financial criteria. In this work, we have tried to address these problems. In order to illustrate the suitability and applicability of the proposed investment decision making model, an empirical study on a set of US domiciled equity mutual funds is carried out. Keywords: Compromise Programming, Equity Mutual Funds, Multi-Criteria Decision Making, Portfolio Selection, Socially Responsible Investment. 55 MO4B (invited) A Conic Scalarization Method in Multi-Objective Optimization Kasimbeyli, Refail (*) Anadolu University This paper presents the conic scalarization method for scalarization of nonlinear multiobjective optimization problems. We introduce a special class of monotonically increasing sublinear scalarizing functions and show that the zero sublevel set of every function from this class is a convex closed and pointed cone which contains the negative ordering cone. We introduce the notion of a separable cone and show that two closed cones (one of them is separable) having only the vertex in common can be separated by a zero sublevel set of some function from this class. It is shown that the scalar optimization problem constructed by using these functions enables to characterize the complete set of efficient and properly efficient solutions of multi-objective optimization problems without convexity and boundedness conditions. The conic scalarization method guarantees a most preferred efficient or properly efficient solution, if a suitable scalarizing parameter set consisting of a weighting vector, an augmentation parameter, and a reference point is chosen. Keywords: Achievement Scalarizing Functions, Multi-Objective Optimization, NonLinear Multi-Objective Optimization, Scalarization, Weighted Sum. 56 MO3B (invited) A Decision Model for Identifying and Prioritizing the Capability Gaps in Defense Planning Kandakoglu, Ahmet (*) Turkish Air Force Command Kahraman, Y.Rıza Turkish Air Force Command Topcu, Ilker Istanbul Technical University Capability-Based Planning (CBP) is a systematic approach to force development that aims to propose the most appropriate force options to meet the defense priorities. The starting point of this approach is to identify and prioritize the requirements and gaps in operational military capabilities. The more accurately these requirements and gaps are quantified, the better force options are developed and resources are distributed. Therefore, the decision-makers require analytical tools based on quantitative methodologies for managing this critical and highly complex process in an effective and timely manner. In this study, the aforementioned problem is extended to the multiple criteria decision-making (MCDM) environment in order to be systematically analyzed and quantified. A decision model based on Analytic Hierarchy Process (AHP) method is developed for identifying and prioritizing the capability gaps. In the model, first, the capability requirements with respect to possible scenarios are identified. The AHP method is applied to weight these capabilities from the subjective judgments of the experts. Then, the capability gaps are assessed between the identified requirements and the current or programmed force capabilities. Finally, the gaps are prioritized using the weights of the capabilities to provide valuable insights for the planning actions to fulfill these gaps in the CBP process. Keywords: Analytic Hierarchy Process, Capability Based Planning, Multi-Criteria Decision Analysis. 57 WE1F (invited) A Decision Support System for Fuzzy Portfolio Selection Ivorra, Carlos (*) Calvo, Clara Liern, Vicente Universidad de Valencia We present a Decision Support System for fuzzy portfolio selection with cardinality and semi-continuous variable constraints based on a multi-criteria approach. It is intended for investors interested in socially responsible investment that assume as main criterion an adequate trade-off between expected return and risk but they are willing to consider alternative similar solutions taking into account several diversification or non-profit goals. To this end, we present a fuzzy efficient frontier to the investor where he or she can see the best options in a selected zone as well as the consequences of slight movements with regard to the various criteria involved. Keywords: Fuzzy Mathematical Programming, Multi-Criteria Decision Making, Portfolio Selection. 58 TU5B (contributed) A Framework for Sustainable for Third Party Logistic Supplier Identification Karpak, Birsen (*) Youngstown State University Recent research (among 40 large 3PL companies) illustrated that 3PL suppliers are committed themselves to environmental sustainability goals. 3PL customers have also shown increasing interest in the environmental sustainability capabilities of the 3PL suppliers. The choice of location such as offshore vs. onshore, close to the customer vs. close to the raw material source as well as the choice of mode of transportation can have a significant impact on the GHG emissions. Carbon footprint receives the most attention. Social aspects and their integration into business seems to be much less advanced and therefore even more difficult to trace than environmental aspects. The economic dimension was regarded as most important, stemming from the argument that without economic success, there would be no supply chains. A Multiple criteria model has been proposed for 3PL suppliers considering economic, environmental and social dimensions of sustainability. Criteria as well as potential 3PL suppliers will be prioritized with an illustrative example. Keywords: 3PL, Analytic Network Process, Multi-Criteria Decision Making, Sustainability, Sustainable Supply Chains. 59 FR3C (contributed) A Fuzzy Model for Project Portfolio Selection P´erez, F´atima (*) Caballero Fern´andez, Rafael University of M´alaga G´omez, Trinidad Liern, Vicente Universidad de Valencia Decision makers usually have to face a budget and other types of constraints when they have to decide which projects are going to be undertaken (to satisfy their requirements and guarantee profitable growth). Our purpose is to assist them in the task of selecting projects portfolios. We have approached this problem by proposing a general nonlinear binary multi-objective mathematical model, which takes into account all the most important factors mentioned in the literature related with Project Portfolio Selection and Scheduling. Due to the existence of uncertainty in different aspects involved in the aforementioned decision task, we have also incorporated into the model some fuzzy parameters, which allow to represent information not fully known by the decision maker/s. The resulting problem is both fuzzy and multiobjective. The results are complemented with graphical tools, which show the usefulness of the proposed model to assist the decision maker/s. Keywords: Fuzzy Mathematical Programming, Multi-Objective Optimization, Project Portfolio Selection. 60 WE2C (contributed) A Fuzzy Numbers Regression for Inducing OECD Wellbeing to WCY Competitiveness 2012 Wen, Hsin-Hui (*) Chung Hua University Ko, Yu-Chien Chung Hua University Tzeng, Gwo-Hshiung Kainan University The better life could be the basics of the national competitiveness. In order to verify this assumption, an induction of crossing systems is proposed to illustrate the conditional dependence of better life on competitiveness. The technique of crossing systems is built on fuzzy numbers that associate criteria values to decision objects. Based on a homogeneous–scaling of the fuzzy numbers, a classification regression can make inductions of crossing systems available. For an illustration instance, Organization for Economic Co-operation and Development (OECD) better life and World Competitiveness Yearbook (WCY) competitiveness in 2012 are used as dataset, dominance-based rough set approach (DRSA) is designed to give the fuzzy numbers. The result reveals that the better life does not have significant evidences to support national competitiveness; contrarily, the national competitiveness has the regression evidence on the better life effectively. The weights of the regression further help users to realize the priorities of the aggregated competitiveness. Keywords: Competitiveness, Fuzzy Numbers, Fuzzy Set Extension Of DRSA, Wellbeing. 61 TH4F (contributed) A Genetic Algorithm Approach for Solving Multi-Objective Factory Layout Problem Ho, Phong (*) Vietnam National University-HCMC Huy, Truong Quoc, Ho Vietnam National University-HCM Facility layout is a significant decision which is associated with manufacturing system performance. Traditionally, approaches to layout design attempted to minimize the total transport distance among the facilities. Recently, the facility layout is viewed as a multi-objective optimization problem to meet the competitive environment. However, the existing works have not paid an attention to the flexibility of the manufacturing system. The research is aimed at using multi-objective optimization in the context of a factory layout problem with new proposed criterion namely reachability, an element used to evaluate the ability of the system to reach to another facility when the current facility is out of order. The optimization model is built based on Genetic Algorithms (GAs) to solve the factory layout problem which is modeled as a multi-objective optimization problem – minimizing total travel time, maximizing total closeness rating scores, and maximizing total layout flexibility. A numerical example is presented to demonstrate the effectiveness of GAs approach as well as the application of the reachability criterion in facility layout problems. Keywords: Factory Layout, Genetic Algorithms, Multi-Objective Optimization, Reachability. 62 TH4F (contributed) A Genetic Algorithm for a Multi-Objective Parallel Machine Scheduling Problem with Special Process Constraints Sara¸c, Tugba (*) Aydinbeyli, Yakup Emre Ko¸canli, M¨ uzeyyen Melek Eskisehir Osmangazi University In this study, a special type of identical parallel machine scheduling problem with n products and m identical machines is considered to determine that the products should be produced in which machine and order in a biscuit factory. The multi-objective goal programming model with special process constraints is proposed for this problem. Since size of this kind of problems is usually large in real life, they cannot be solved with exact solution methods. Therefore, a genetic algorithm with an original crossover operator is developed for solving this problem. The effectiveness of the developed solution approach is tested on small, medium and large size randomly generated test instances. Finally, obtained results for the real-life problem are discussed. Keywords: Scheduling. 63 TH2D (invited) A Group Decision Making Model Based in Goal Programming with Fuzzy Hierarchies. An Application to Regional Forest Planning Jim´enez L´opez, Mariano (*) Universidad del Pa´ıs Vasco Arenas-Parra, Mar Bilbao, Amelia Rodr´ıguez-Ur´ıa, Mar´ıa Victoria Universidad de Oviedo A multi-criteria group decision making model is presented. In this model the relative importance between the different decision makers in the group is imprecise. To address this problem a goal programming (GP) approach is used. We suppose that each decision maker has its preferred solution, obtained applying any multi-objective programming technique (compromise programming, weighted GP, lexicographic GP, extended lexicographic GP, GP with fuzzy hierarchies, ...). These solutions are used as targets in an aggregated GP model in which the normalized unwanted deviations are interpreted in terms of degree of achievement of the targets. Using this approach fuzzy binary relations are employed to model the imprecise hierarchy between decision makers. This way a GP model with fuzzy hierarchies (FHGP) is built. The solution of this model is proposed as a consensus decision. To show the applicability of our proposal a regional forest planning problem is tackled. The objective is determining tree species composition in order to improve the values achieved by Pan-European sustainable indicators. This problem involves competing uses and stakeholders which, regarding the aforementioned indicators, have a different preference scheme. In the first place a decision model for each stakeholder is solved. Therefore their corresponding preferred plan is obtained. These preferred plans are aggregated in a GP model in which the fuzzy hierarchy between stakeholders is incorporated. The solution of this FHGP model is offered to the stakeholders as a consensus plan. Keywords: Decision Maker’s Preferences, Fuzzy Sets, Goal Programming, Group Decisions Making. 64 TU4B (contributed) A Hybrid Dynamic MCDM Model-Tourism Competitiveness Improvement Strategy Tzeng, Gwo-Hshiung (*) Lu, Ming-Tsang Kainan University Peng, Kua-Hsin Peng, Hsin-Chuan Yuanpei University Multiple Criteria Decision Making (MCDM) is scientific analytical method that can help decision-makers for evaluating and selecting the best alternative based on multiple criteria. However, traditional MCDM have some limitations/defects for solving the real-world problems. First, conventional MCDM assumes independent criteria with a hierarchical structure. Second, conventional MCDM only obtains relatively good solutions from existing alternatives. Third, conventional MCDM merely allows the selection and ranking of alternatives or strategies. Therefore, a Hybrid Dynamic MCDM combined with DANP (DEMATEL-based ANP) method must be developed to overcome the defects of the traditional MCDM method. Finally, an empirical case is presented to demonstrate the effectiveness of a Hybrid Dynamic MCDM model combining DANP and VIKOR for evaluating tourism competitiveness of four Asian countries, namely Japan, Malaysia, Singapore and Taiwan to identify competitiveness gaps and explore strategies for improving tourism competitiveness. Decision makers should increase the priority of the cause criteria based on the influential relation map in advance, and to successfully create a high tourism competitiveness to achieve the aspired/desired levels. Keywords: Analytic Network Process, Competitiveness, DEMATEL, Hybrid Approaches, Multi-Criteria Decision Making, Multi-Criteria Evaluation. 65 TU5D (invited) A Hybrid Method for Supplier Order Quantity Allocation Kuo, R. J. (*) Bai, C. M. Hu, T. L. Taiwan University of Science and Tech. With fiercely global competition, most of manufacturing functions in supply chain have become more and more important. Especially materials purchasing, it is a critical link between outsourcing and manufacturing. Thus, an efficient purchasing can continuously help enterprise enhance efficiency, reduce cost and increase profit. However, in consideration of risk diversification, the current manufacturing considers multiple source suppliers. In selection of key suppliers, most enterprises attempt to obtain the best product quality and maximize interest of supply chain partners. Thus, selecting suitable suppliers can provide high-quality product and reduce total costs for the enterprises. Thus, supplier selection and order quantity allocation are still two important issues for the current manufacturing sector. Therefore, this study intends to propose a two-stage approach for order allocation. In the first stage, one of the association rule mining techniques, TD-FP-growth algorithm, is employed to select important suppliers from the existing suppliers and determine the importance of each supplier. Then, a hybrid of artificial immune network (Opt-aiNet) and particle swarm optimization (PSO) (aiNet-PSO) is proposed to allocate the order quantity for the key suppliers with minimum cost by considering multiple criteria including purchasing cost, ordering cost, delivery cost, and defective material cost. In order to verify the proposed two-stage approach, a case company’s daily purchasing ledger focusing on the consumer electronic product manufacturers is employed. The computational results indicate that TD-FP-growth algorithm is able to select the key suppliers using the historical data. Besides, the proposed hybrid method, aiNet-PSO, can provide lowest cost compared to those of genetic algorithm, PSO, and artificial immune system. Thus, the proposed approach is able to help case company reduce the outsourcing cost. Keywords: Allocation, Data Mining, Decision Support Systems, Evolutionary Computations, Hybrid Approaches, Supply Chain Management. 66 TU5C (contributed) A MAUT Approach for the Management and Remediation of a Site Contaminated by Uranium Processing Activities Mart´ın, Miguel C. (*) Jim´enez, Antonio Mateos, Alfonso Technical University of Madrid P´erez-S´anchez, Danyl CIEMAT Pridneprovsky Chemical Plant, located at Dneprodzerzhinsk (Ukraine), was one of the largest uranium processing enterprises of the former USSR, producing a huge amount of uranium residues stored in 9 tailing impoundments and contaminating different buildings. The processing of uranium stopped in 1990, due to the disintegration of the USSR, and since then financial resources were not put in remediation programs until 2003. Nowadays, more than 20 enterprises not related to uranium processing activities are in operation at this territory and a residential area is located rather close to the zone. Since 2003, the Government of Ukraine began a systematic investigation of this territory. Environment monitoring and site surveillance programs were established in 2005. During 2008-09, preliminary safety assessment for the identified hazards was been carried out and the resulting strategies are being implemented during the State Remediation Programme (2010-14), for which an international consortium of experts was created within the framework of ENSURE projects. Among the remediation problems that still are waiting for optimal strategy to be chosen are the management and decommissioning of the high contaminated buildings, the largest tailing at Dnieprovskoe, which is partly covered with thick layer of phosphogypsum and also the wet uranium tailing at Suhachevskoe, still partly covered with water. The situations described above are complex decision-making problems in which several economical, social and environmental conflicting objectives must be taken into account simultaneously, being necessary a formal analysis. We have used the GMAA DSS based on an additive multi-attribute utility model to deal with these problems. It accounts for uncertainty about the strategy impacts and admits incomplete information about the decision-makers’ preferences. The system provides different sensitivity analyses that take advantage of the imprecise inputs to reach a final recommendation. Keywords: Applications of MCDM, Multi-Attribute Utility Theory, Restoration, Uranium Processing. 67 WE2D (invited) A MCDA Model to Support Public Safety Management Policy Mota, Caroline (*) Gurgel, Andr´e Unoversidade Federal de Pernambuco Public safety planning is a subject of great interest in the modern society. This paper contributes with an strategic issue for public safety management, regarding the designation of different public policies according to different areas in one region. Therefore, this article proposes a multicriteria decision model that seeks to prioritize areas for a given region using demographic and socioeconomic issues. Consequently, it’s possible to perform different strategies in a region according to a set of factors that can influence criminality. The model was applied in a Brazilian city, where it was divided into human development units and resulting in 62 different zones of criminality. Keywords: Multi-Criteria Decision Making, Outranking Methods, Public Safety. 68 MO3D (contributed) A Model for a Spatial Forest Problem Taking into Account Economic and Environmental Objectives Hernandez, Monica (*) University of Malaga G´omez, Trinidad University of M´alaga Caballero Fern´andez, Rafael University of Malaga Leon, M. Amparo University of Pinar del Rio Molina Luque, Juli´an University of M´alaga In this work, a forest harvesting planning problem is solved taking into account economic factors of timber production, environmental protection and other aspects related to. The forest problem is formalized via a non linear multiobjective programming model, in which carbon captured is considered, and taking into account spatial constraints aimed at limiting the maximum adjacent area to which clearcutting can be applied. The final outcome is a highly complex problem that is solved by applying a metaheuristic method. The model proposed is applied to a real timber production plantation in Cuba and also to a series of hypothetical forests, randomly generated. Keywords: Forest Management, Fractional Programming, Harvest Scheduling, MultiObjective Optimization. 69 MO5H (poster) A Multi Criteria Decision Analysis-Based Methodology for the Risk Assessment of Flood Hazards at the Regional Scale Isigonis, Panagiotis (*) Critto, Andrea Gallina, Valentina Marcomini, Antonio Semenzin, Elena Torresan, Silvia Zabeo, Alex Ca’Foscari University of Venice In recent years, the frequency of water-related disasters has increased and flood events in Europe (e.g. 2002 in central Europe, 2007 in UK, 2010 in Italy) have caused physical, environmental and socio-economic damages. The current European legislation (e.g. Floods Directive) enforces the establishment of frameworks for the assessment and management of flood risks and aims towards the reduction of the consequences for human health, the environment, the cultural heritage and the economic activity associated with floods in the European Union. In this context, within the KULTURisk project (FP7), a Regional Risk Assessment (RRA) methodology is proposed to evaluate the benefits of risk prevention in terms of reduced environmental risks due to floods, through a GIS-based tool that produces maps of the total risk index for the regions under analysis. The methodology is based on the KULTURisk framework and allows the identification and prioritization of targets (i.e. people, buildings, infrastructures, agriculture, natural and semi-natural systems, cultural heritages) and areas at risk from floods in the considered region by comparing the baseline scenario (i.e. current state) with alternative scenarios (i.e. where different structural and/or non-structural measures are planned). The RRA methodology employs Multi-Criteria Decision Analysis (MCDA) methods in order to integrate stakeholder preferences and experts judgements into the analysis as well as for the calculation of the total risk index. The RRA methodology is flexible and can be adapted to different case studies (i.e. large rivers, alpine/mountain catchments, urban areas and coastal areas) and spatial scales (i.e. from the large river to the urban scale). Overall, the methodology intends to help decision-makers in examining the possible environmental risks associated with uncertain future flood hazards and in identifying which prevention scenario could be the most suitable one. Keywords: Applications of MCDM, Environmental Decision Making, Risk Assessment. 70 TU4C (contributed) A Multi-Criteria Aiding Tool for the Assessment of French Statutory Urban Planning based on ELECTRE TRI. A Case Study of the Local Plan of Toulouse, France Prevost, Aur´elie (*) Molines, Nathalie Universit´e Technologie de Compi`egne Bandet, Jean DGDUD / Toulouse M´etropole French legislation, according to European legislation (Directive 2001/42 CE), recently provided an environmental assessment (including ex-post assessment) for some French local plans. The ex-post assessment requires building a set of “environmental indicators”. Nevertheless, those indicators don’t allow the planners to understand the traditional dynamics of the plan regulations, which are complex and can participate at the same time in the implementation of “urban sustainability principles” lead by the legislation. The city of Toulouse engaged a PhD research in order to propose methodological insights for the evaluation of the impacts of the plan’s regulations. The research focuses on 3 main issues of traditional French statutory planning: constructability, the relation between built and unbuilt surfaces, and the amenities of each statutory zone. A set of indicators is created for the different scales of the plot, the statutory zone and its geographical units. Later, the indicators are used to highlight the existing links between the characteristics of the territory and the regulation. After the creation of the indicators, the method proceeds in two steps: 1. A comparison of each zone and deduction of the regulation’s effects, thanks to the indicators. 2. Positioning of each zone according to the characteristics of the whole territory. Here, we propose to classify each zone in accordance to different objectives to achieve. This third step is divided into two aims: an ex-post assessment of the plan’s effect on morphological criteria (the building morphology) and amenities in the zone, and second, an ex-ante assessment to help planners find the adequate change for the definition of zones and regulations. For this step, GIS software and the ELECTRE TRI method are used. This paper proposes to present the second step of the method. After a synthetic description of the project, we will focus on the implementation of the ELECTRE TRI method, and the results obtained. Keywords: Amenities, Characterization Of Statutory Zones, Constructability, Decision Support, ELECTRE TRI, Impacts Of Statutory Planning. 71 WE2C (contributed) A Multi-Criteria Decision Analysis Methodology based on Multi-Attribute Value Theory and Fuzzy Logic for Quantitatively Scoring the Reliability of Ecotoxicological Data Isigonis, Panagiotis (*) Critto, Andrea Marcomini, Antonio Semenzin, Elena Zabeo, Alex Ca Foscari University of Venice Ciffroy, Philippe EDF The advances in the field of ecotoxicology suggest exploring new ways of applying the current methods for the evaluation of toxicity as well as identifying ways to make more efficient use of the existing ecotoxicological datasets. In this scope and context, a new methodology is presented which intends to assist the optimization of existing methods, by providing a tool for assessing the relevance and reliability of ecotoxicological data for the definition of Species Sensitivity Distributions (SSDs), within the framework of ecological risk assessments (ERA). In order to estimate a single aggregated reliability score for a given ecotoxicological datum, a ‘Multi-Criteria Decision Analysis (MCDA)-based’ Weight of Evidence (WoE) methodology has been developed, including a hierarchical structure of 57 evaluation criteria, which was created based on the review of the state of the art frameworks for the assessment of ecotoxicological data. The methodology is based on the Multi-Attribute Value Theory (MAVT) and is able to integrate different types of inputs. It incorporates the use of Fuzzy Logic operators for handling the inherent uncertainty, which appears in the form of unreported information as well as possible lack of knowledge of the experts. A panel of scientific experts on ecotoxicology was involved throughout the development of the methodology for identifying, if any, the relations between criteria and evaluating the hierarchical structure to be used in the aggregation process. Expert Knowledge is incorporated in the methodology and was extracted from the panel with the use of direct techniques, i.e. questionnaires and dedicated meetings. The methodology is planned to be used and tested in case studies, as part of the evaluation and follow up process of the research project “AMORE” (Multi-Criteria Analysis for the Development of a Decision Support Tool for the prevention of Environmental Risks), funded by the National French Research Academy (ANR). Keywords: Fuzzy Logic, Group Decision Making, Multi-Attribute Value Theory, Multi-Criteria Decision Analysis. 72 MO2G (invited) A Multi-dimensional Corporate Social Performance Versus a Multi-dimensional Corporate Social Irresponsibility Shalchian, Homayoon (*) Laurentian University El badraoui, Khalid University of Rennes Bouslah, Kais Andrews University Lilti, Jean-Jacques University of Rennes M’Zali, Bouchra UQAM This study investigates, on one hand, the relation between corporate social performance (CSP) and corporate financial performance (CFP), and on the other hand, the relation between corporate social irresponsibility (CSI) and corporate financial performance. In our previous study, assuming that investors consider and grant the same level of importance to each and all dimensions of CSP, we examined the relation social-financial performance using various performance models. Our previous study also showed that the nature of activities of the firms can have a significant impact on the relation social-financial performance. Our first contribution in this study consists of analyzing the relation of different dimensions of corporate social performance and financial performance separately in different industries. Our second contribution consists of analyzing the relation of corporate social irresponsibility and financial performance in order to compare this relation with social-financial performance. To do this we construct mutually exclusive portfolios based on ’Strengths’ and ’Weakness’ rated by KLD. Our results show that in all industries considered in this study and based on both strength and weaknesses rated by KLD, the high ranked portfolios outperformed the low ranked portfolios. However, the gap between portfolios performances is larger for the dimension ’environment’ in the industry where the nature of activities corresponds to the industry ’Mining and chemicals’. Moreover, our results show that in the industries ’Manufacturing’ and ’Services’ the dimensions ’Employees welfare’ constitutes a stronger signal relative to the other dimensions. Keywords: Corporate Social Performance, Financial Performance, KLD Rating, Screening. 73 WE2D (invited) A Multi-Objective Genetic Algorithm for Integrated Scheduling of Production and Maintenance Activities for a Single Machine Ferreira, Rodrigo (*) Almeida-Filho, Adiel Cavalcante, Cristiano Federal University of Pernambuco This paper proposes a multi-objective model to deal with the integrated problem of scheduling of production and maintenance activities in a single machine setting. Typically, the main issue to be addressed in scheduling problems is to define the sequence in which production activities must be performed in order to maximize one objective function. In addition to production activities, the model proposed in this paper considers the scheduling of maintenance activities in an integrated manner with two conflicting objectives related to minimizing expected tardiness and maintenance costs. Since the complexity of the solution space for this combinatorial problem is high, a multi-objective genetic algorithm is proposed to find solutions that belong to the Pareto frontier, based on the concept of dominance. The results emphasize that the model and algorithm can provide non-dominated solutions promptly. Keywords: Genetic Algorithms, Multi-Objective Optimization, Preventive Maintenance, Production Scheduling. 74 MO2D (contributed) A Multi-Objective Route Choice Model for Health-Conscious Cyclists Wang, Judith Y. T. (*) University of Leeds Cheung, Alan K. L. The University of Auckland Dirks, Kim N. The University of Auckland Ehrgott, Matthias Lancaster University Pearce, Jon The University of Auckland Cyclists form the most vulnerable road user group in terms of injury from traffic accidents as well as exposure to traffic-related air pollution. Ironically, commuter cyclists are often motivated by improvement in health and fitness. Cycleways away from traffic with lower concentrations of pollutants from motorised vehicles are often longer and hence require longer travel times, while alternative routes sharing the road with other traffic, sometimes with buses, might result in exposure to higher pollutant concentrations. To help commuter cyclists achieve their objectives of getting to work on time and maximising their health benefits, we propose a multi-objective route choice model, for example, with minimisation of travel time and pollutants uptake as the two objectives in a bi-objective routing problem. A transport network information database is first constructed with comprehensive information on link type, lane width, gradient, link average speed, traffic volume, etc. such that both travel time and pollutants uptake can be estimated at a reasonably accurate level. In particular, the pollutants uptake will be dependent on exercise level as well as the concentration of pollutants. Given an origin and a destination, to be provided by a cyclist, we apply a multi-objective shortest path algorithm to determine an efficient set of routes such that neither the total travel time nor the total pollutants uptake can be reduced without worsening the other. Profiles of this route choice set in terms of other useful information such as elevation and pollutant concentrations along the route can also be provided. With our model, cyclists can more easily trade off between time and pollutant uptake, and select a route with reasonable travel time at an appropriate exercise level with minimal negative impact on health. In cities with a hilly terrain like Auckland, New Zealand, such information will be extremely valuable to current and potential cyclists. Keywords: Environment, Routing, Shortest Path, Transportation. 75 TU3D (contributed) A Multi-objective Waste Collection Problem ´ Hern´andez-D´ıaz, Alfredo G. (*) Hinojosa, Miguel Angel L´opez-S´anchez, Ana Dolores Pablo de Olavide University Duarte, Abraham Gort´azar, Francisco Rey Juan Carlos A new multi-objective GRASP algorithm has been implemented to solve a waste collection problem. It has been tested over a set of medium sized instances as well as the real instance. The real world problem addressed is the waste collection problem in Seville (Spain). Since vehicles can be only loaded from their right side, a directed graph with two distinguished nodes namely the depot (where vehicles are stored) and the landfill (where vehicles are emptied) has been considered. Each vehicle must visit streets until its capacity is reached several times during the workday. Thus, each vehicle should perform a route combining two or more trips. In the initial trip, vehicles start at the depot and end at the landfill while the rest of the trips start and end in the landfill. The objectives are minimizing the number of vehicles, the total distance and also the longest route with the aim of balancing workloads. The research has been supported by the local public cleaning firm of the city. Keywords: GRASP Algorithm, Routing, Waste Collection Problem. 76 TH3G (invited) A Multicriteria Decision Aiding Methodology to Support Public Administration on Sustainable Energy Action Plans Norese, Maria Franca (*) Dall’O, Giuliano Politecnico di Torino Novello, Chiara The Sustainable Energy Action Plan (SEAP) represents a strategic tool for achieving the greenhouse gas reductions to 2020 for the municipalities that have joined the Covenant of Mayors promoted by the European Commission. The intervention scenarios in relation to the energy retrofits actions in the residential building stocks, that in small-medium municipalities are responsible for more than 60% of CO2 emissions, are normally decided on the basis of an economic analysis that calculates the ratio cost/performance. This type of analysis, however, does not take into account important aspects for small and medium-sized communities such as social aspects, environmental impacts, local economic development and employment. A more comprehensive and effective tool to support the choices of public administrators could be a multiple criteria model. A methodology has been developed to support Public Administration on programming Sustainable Energy Action Plans with a more targeted approach to sustainability. The methodology, that includes the ELECTRE III method, was applied to a medium-size Municipality in Lombardy region (Italy) and the results were compared with the classical economical approach and analyzed in terms of model consistency and adequacy of the methodology to the actual processes of policymaking, planning and action implementation. Difficulties and opportunities that characterize the modeling process and the application of an ELECTRE method in this context will be described in relation to the impacts on the decision processes and the different possible decision aiding interventions. Keywords: ELECTRE Methods, Multi-Criteria Decision Aid, Public Policy, Strategic Planning, Sustainability. 77 TU3C (contributed) A Multicriteria Optimization Model for a Humanitarian Logistics Problem: An Integral Approach Gayt´an-Iniestra, Juan (*) Universidad Aut´onoma de M´exico Mej´ıa-Argueta, Christopher ITESM Campus Toluca Caballero Fern´andez, Rafael University of Malaga Molina Luque, Juli´an University of M´alaga Vitoriano, Bego˜ na Universidad Complutense de Madrid Disasters are phenomena which strike countries around the world. This work introduces an integral optimization model to prepare: Distribution, evacuation, location of facilities and meeting points, as well as a stock management policy during floods with multicriteria (total cost, as well as equity minimizing the maximum evacuation and distribution flowtimes). The efficient frontier is built through the weighthing and the epsilon-constraint methods through the use of commercial software. The usefulness and the robustness of the model is validated through numerical experiments on 11 different instances including a mexican case study. Keywords: Applications of MCDM, Case Study, Humanitarian Logistics, Multi-Objective Optimization, Pareto Optimal Solutions, Sensitivity Analysis. 78 FR3E (contributed) A Multicriteria Prediction Model for Project Classifications Laryea, Rueben (*) Stockholm University Project distress predictions are essential in project management. Developing appropriate methods to classify projects and building prediction models for Multi-Criteria decisions requires empirical methods to minimize misclassification errors. This paper carries out Multi-criteria analysis to classify projects using a preference disaggregation method, UTADIS. The UTADIS requires a predefined classification which is implemented using critical path analysis. The methods are applied on three projects and results in a no misclassification error and an effective prediction model. Keywords: Classification, Critical Paths, Multi-Criteria Decision Model, Preference Disaggregation, Project Risk. 79 TH2F (contributed) A Multiple Objective View on Robust Optimization and Stochastic Programming with Finite Scenario Sets Klamroth, Kathrin (*) University of Wuppertal K¨obis, Elisabeth University Halle-Wittenberg Sch¨obel, Anita Georg-August University G¨ottingen Tammer, Christiane University Halle-Wittenberg Many real world optimization problems are contaminated with uncertain data. One way of dealing with such optimization problems is described in the concept of robustness: Instead of assuming that all data are known, one allows different scenarios for the input parameters and looks for a solution that works well in every uncertain scenario. Another prominent way of dealing with uncertain optimization is the field of stochastic programming. Different from robust optimization, stochastic programming assumes some knowledge about the probability distribution of the uncertain data. The objective usually is to find a feasible solution (or a solution that is feasible with a certain probability) that optimizes the expected value of some objective or cost function. In this talk, we will link different concepts of robustness and of stochastic programming that usually have been considered fundamentally different through the formulation of deterministic multiple objective counterparts, where objective functions are defined by different scenarios of the uncertain model. Assuming that the set of scenarios is finite, we will show that many uncertain optimization problems have a deterministic counterpart in this framework. Similarly, we show that many different concepts of robustness and of stochastic programming can be described as special cases of a general nonlinear scalarization method by choosing the involved parameters and sets appropriately. This forms the basis for an analysis of the interrelations and also the differences between established concepts in robust optimization and stochastic programming. By providing additional trade-off information between alternative efficient solutions, the multiple objective counterpart can facilitate the decision making process when deciding for a most preferred robust solution. Keywords: Non-Linear Scalarization, Robust Optimization, Stochastic Programming, Uncertainty. 80 TU4F (contributed) A Neural Network based Hybrid Evolutionary Multiobjective Optimization Algorithm for Computationally Demanding Problems Hakanen, Jussi (*) Kokko, Tommi Sindhya, Karthik University of Jyv¨askyl¨a Solving real-world multiobjective optimization problems (MOPs) is a computationally demanding task due to, e.g., more accurate and complex simulation models used and the number of objective functions considered. One approach to tackle this challenge is to utilize computationally inexpensive surrogate functions to approximate the objective functions. Hybrid evolutionary multiobjective optimization (EMO) methods are also used to handle computationally demanding MOPs. Here a global search (EMO method) and a local search are combined to improve the performance of EMO methods. In local search we optimize an achievement scalarizing function using a suitable optimization method and is sparsely applied on selected solutions. In this research, we utilize the benefits of surrogate functions and hybrid EMO methods. Here, we consider neural networks (NNs) as an example of surrogate functions and use them to approximate the objective functions. The surrogate functions are built using a small set of sample points evaluated with the original computationally costly model. We combine a hybrid EMO method with a NN in a way that the objective function evaluations of EMO part are replaced with a NN and the objective function evaluations of the local search are performed with the original model. This leads to a significant saving in the computation time since most of the evaluations of hybrid EMO methods are used in the EMO part. Here we consider multilayer perceptron and a radial basis network as examples of NNs. We test the efficacy of our combined approach using a computationally expensive engineering problem. Keywords: Computational Cost, Evolutionary Multi-Objective Optimization, Hybrid Approaches, Surrogate Modeling. 81 TH3D (contributed) A New Dominance Measuring Method for MCDM with Ordinal Information About DM’s Preferences Aguayo, Ernesto (*) Jim´enez, Antonio Mateos, Alfonso Technical University of Madrid Different methods based on dominance intensity measures have been recently proposed by different authors to deal with imprecision concerning MCDM problems. These methods compute dominating and dominated measures on the basis of the pairwise dominance values, which are then used in different ways to derive dominance intensity values on which the ranking of alternatives is based. In this paper we propose an extension of a dominance measuring method previously published by the authors to manage MCDM problems in which the decision maker’ preferences are represented by ordinal information, i.e a ranking of criteria is provided based on their relative importance, whereas alternatives are also ranked in each criterion. First, optimization problems are solved to derive pairwise dominance values using the centroid function to represent the relative importance of criteria, while ordinal information about alternative in each attribute is incorporated as constraints. Once the dominance matrix is computed, we ponder the elements of the dominance matrix with the distance that exists between the mean weight vector, which is the result of each mean value from the end points of the polytope that represents the imprecision of the problem, and the optimal weight vector, that is the result of the optimization model for that element. Finally, a dominance intensity measure is computed from the pondered dominance matrix on which the ranking of alternatives is based. Monte Carlo Simulation techniques are used to compare the performance of the extension we proposed with other dominance measuring methods, with an approach proposed by Sarabando and Dias for MCDM problems with ordinal information. Two measures of efficacy are considered, the proportion of all cases in which the method selects the same best alternative as in the TRUE ranking (hit ratio) and how similar the overall alternative-ranking structures are in the TRUE and the method-driven rankings (rank-order correlation). Keywords: Dominance Measuring Methods, Imprecise Information, Monte Carlo Simulation, Multi-Criteria Decision Making. 82 TU3F (invited) A New Evolutionary Algorithm for the Bi-objective Ring Star Problem Gal´e, Carmen (*) Calvete, Herminia I. Iranzo, Jos´e A. Universidad de Zaragoza The ring star problem aims to locate a cycle through a subset of nodes of a graph and assign each non-visited node to a visited one. The goal is to minimize the ring costs and the assignment costs. The bi-objective ring star problem (B-RSP) arises when both objectives are considered individually instead of jointly. In this paper, a new evolutionary algorithm is proposed to approximate the set of Pareto optimal solutions. Previous evolutionary algorithms for solving the B-RSP use the random keys encoding mechanism, which easily allows to applying evolutionary operators. With this encoding at hand, the chromosome contains all the information about the cycle. Hence, in order to evaluate the fitness of a chromosome, the algorithm directly computes the cost of the edges in the cycle and the cost of the assignments. The distinctive aspect of the proposed algorithm is that the chromosome only contains information on the nodes in the cycle, but not on their position. Hence, we encode each chromosome as a binary vector of dimension the number of nodes in the network, where 1 indicates that the node is in the cycle. After knowing the nodes in the cycle, the complete description of the cycle is obtained by solving a TSP on these nodes. The algorithm proceeds by applying uniform crossover operator, mutation and local search. In spite of the more complex process to compute the cycle, the algorithm has shown its efficiency in terms of the quality of the Pareto frontier approximation provided and the computing time when applied to a set of benchmark problems. Keywords: Bi-Objective Ring Star Problem, Evolutionary Algorithm. 83 FR3C (contributed) A New Exact Method for Optimizing a Linear Function over an Integer Efficient Set Chaabane, Djamal (*) USTHB During these two decades, researchers and practitioners have been increasingly interested by the problem of optimizing a linear function over the efficient set of multiple objective linear programming problem . Several methods and algorithmic ideas have been developed, in general, these approaches can be classified and grouped according to the methodological concepts. Among others, adjacent vertex search technique, face search algorithm, algorithms include Branch and bound, Lagrangian relaxation methods, dual approach and other methods. An overview of these approaches can be found in Yamamoto (2002). Contrary to the continuous case, few exact algorithms have been suggested for solving the problem involving discrete decision variables. The first attempt for optimizing over the integer efficient set is due to Nguyen 1992), where only an upper bound value for the main objective is proposed. Thefirst exact method was developed by Chaabane and Abbas (2006) based on a simple selection technique that improves the main objective value at each iteration. Two types of cuts are used and performed successively until the optimal value is obtained and the current truncated region contains no integer feasible solution. In the present paper, we propose a new exact algorithm, using an augmented weighted Tchebychev norm, combined with a reduction process of the admissible region by the successive addition of constraints. Sub-programs that were integrated in the above cited methods are avoided and the technique produces two outputs : an optimal efficient solution and a sub-set of efficient solutions. Keywords: Efficiency, Integer Linear Programming, Multi-Criteria Decision Analysis, Multi-Objective Optimization. 84 MO4D (invited) A Parameterized Approximation for Matroids with Multiple Objectives Tlilane, Lydia (*) Gourv`es, Laurent Monnot, J´erˆome Universit´e Paris Dauphine We consider an optimization problem on a matroid with multiple objectives. One has to find a base that maximizes n additive functions. Taken separately, every objective can be optimized by applying the well known greedy algorithm (also known as Kruskal’s algorithm for maximum weight spanning trees). However, the existence of a feasible solution that maximizes simultaneously all objectives is unlikely. In addition, the existence of a solution that maximizes the function of the least satisfied objective has been shown NP-complete in [E. Ehrgott, On matroids with multiple objectives. Optimization 38.1 (1996): 73-84]. Thus, in order to obtain a satisfactory solution in polynomial time, we seek for an approximation of the ideal point. A feasible solution (base) B is a (e1 , ..., en )approximation of the ideal point if the ratio between the ith value of B and the optimum on objective i is at least ei for all i = 1, . . . , n. There exist conflictual instances which are unbalanced, i.e., every feasible solution is close to optimality for one objective, and very far from optimality for another. In this case, we can not guarantee anything. If the matroid is simple (every element or couple of elements is independent), there exists a (1/2, 1/3)-approximation for n = 2 objectives which is tight [L. Gourv`es, J. Monnot and L. Tlilane. Approximate Tradeoffs on Matroids, ECAI 2012: 360-365]. However, if n ≥ 3, then the problem is not constant-approximate even if the matroid is simple. To go further and circumvent this negative situation, we resort to a set of parameters a1 , . . . , an where ai is the maximum value for an element of the matroid on objective i. Our contribution is an approximation depending on ai defined as a family of functions fn (ai ). We propose a polynomial time algorithm providing a (fn (a1 ), ..., fn (an ))-approximate base, according to the possible values of ai . Keywords: Approximation of The Ideal Point, Matroids. 85 TU5E (invited) A Performance Monitoring System based on a Combined MCDM Methodology Kandakoglu, Ahmet (*) Turkish Air Force Command Kandakoglu, Makbule Turkish Petroleum Corporation This study proposes a performance monitoring system based on the combined utilization of Analytic Hierarchy Process (AHP), Weighted Product (WP), Utility Theory (UT) and TOPSIS (Technique for Order Performance by Similarity to Ideal Solution) methods. In the proposed system, the AHP is applied to identify the key performance criteria, structure the criteria hierarchy and determine the weights of the criteria from the subjective judgments of experts. UT is used to construct the utility functions for each criterion using the organizational standards. WP method is then utilized to calculate the final performance values by penalizing the poor criteria measurements. TOPSIS method is applied to make an evaluation with respect to the standards. Finally, a case study is presented to illustrate the potential and applicability of the proposed system for evaluating the performances of departments and district managements in Turkish Petroleum Corporation. The results show that building this kind of systems enables the aggregation the big data stored in information systems to support the decision making processes within the organization and also allows the decision-makers manage this business processes in an effective and timely manner. Keywords: Analytic Hierarchy Process, Multi-Criteria Decision Analysis, Performance Measurement, TOPSIS, Weighted product. 86 FR2G (contributed) A Preference-Based Facility Location Problem Krupi´ nska, Katarzyna (*) Wroclaw University of Economics We consider the problem of locating facilities on a directed graph. Facilities are allowed to be sit at the vertices of the graph. There is also defined a binary relation on the power set of the set of arcs, according to which preferred paths are determined. In an allocation phase, clients are assigned to facilities based on preferred paths, while in a location phase different location patterns are compared using another preference relation defined on the power set of the set of paths. We formalize the solution concept and present conditions on preference relations under which a prespecified solution may be obtained. Keywords: Location, Preferences. 87 MO4D (invited) A Primal-Dual Approach for Optimizing Ordered Weighted Average in Perfect Matching Problems Weng, Paul (*) LIP6 - UPMC Nguyen, Viet Hung LIP6 - UPMC A perfect matching [Edmonds, 1965] for a (non necessarily bipartite) graph is a subset of non-adjacent edges that cover every vertex. In this work, we assume the graph is weighted and our aim is to find a perfect matching whose edge weights are fairly distributed. This issue is important for instance if one wants to be fair to the pairs of nodes in a matching. The standard formulation of the weighted perfect matching problem, which consists in finding a perfect matching that optimizes the sum of the edge weights, does not allow any control on the distribution of the edge weights. To model fairness, we use the Ordered Weighted Average (OWA) with decreasing OWA weights [Weymark, 1981], instead of the sum, to aggregate the edge weights. While OWA is a non-linear aggregating function, a 0,1-linear program can be formulated for finding an OWA optimal perfect matching, exploiting a linearization of OWA proposed by Ogryczak and Sliwinski (2003). For solving this problem, we propose an approximation scheme with a guaranteed performance based on a primal-dual approach. Keywords: Matching, Multi-Objective Optimization, Ordered Weighted Av. 88 WE1G (contributed) A Semidefinite Programming Approach for Solving Multiobjective Linear Programming Blanco, Victor (*) Universidad de Granada El-Haj Ben-Ali, Safae Universidad de Sevilla Puerto, Justo Universidad de Sevilla Several algorithms are available in the literature for finding the entire set of Pareto-optimal solutions of MultiObjective Linear Programming (MOLP). However, all of them are based on active-set methods (simplex-like approaches). We present a different method, based on a transformation of any MOLP into a unique lifted SemiDefinite Program (SDP), the solutions of which encode the entire set of Pareto-optimal extreme point solutions of MOLP. This SDP problem can be solved, among other algorithms, by interior point methods; thus our approach provides a new non-active-set method to find the set of Pareto-optimal solutions of MOLP. Keywords: Linear Programming, Multi-Objective Optimization, Semidefinite Programming. 89 FR3B (contributed) A Sequential Goal Programming Model for University Timetable Construction Skocdopolova, Veronika (*) University of Economics, Prague Goal programming is a modelling technique usually used for multi-criteria decision making. Timetabling at universities, formulated as an integer or binary programming model, is a problem that belongs to difficult problems. There are two basic approaches for solving the timetabling problem. The first is to solve one complex model usually via integer programming, where often various heuristic and metaheuristic methods (because of computational difficulty) are used. The other approach consists in decomposing the problem into interrelated stages. This paper presents a sequential model, which uses goal programming for solving the timetabling problem in all stages. The model is inspired by model applied by Al-Husain et al. (2011) on scheduling problem at Kuwait University, College of Business Administration. Al-Husain et al. decompose the timetabling problem into three interrelated parts, where each stage is optimally solved and the outputs are used as input in the next stage. At first teachers are assigned to courses, then courses are assigned to time slots, and finally time slots are assigned to classrooms. This approach enables solving the timetabling problem in a reasonable time. The presented model is applied to timetable construction of department of econometrics at University of Economics, Prague. Some new requirements are added contrary to Al-Husain’s model. E.g. there are some courses that have to be in computer classrooms, some of them have to be in a concrete room; some pairs of courses have to be taught simultaneously – one in non-computer classroom, the other one in computer classroom, students rotate every week from one room to the other one, because students do not need PC every lesson of the semester and there are not enough disposable computer classrooms for the department. The aim of the model is to satisfy teacher’s preferences of teaching time. The research is supported by the Internal Grant Agency of the University of Economics (grant no. F4/11/2013). Keywords: Applications of MCDM, Goal Programming, Timetabling. 90 MO4F (contributed) A Stakeholder based Multi-Criteria Evaluation Framework for City Distribution Milan, Lauriane (*) Macharis, Cathy Verlinde, Sara Vrije Universiteit Brussel Urban areas are facing several challenges, both in terms of logistical performance and environmental impact. One of them is how to organise freight transport in a sustainable way. Many innovative city distribution concepts that have been tested fail because not all stakeholders were taken into account in the decision-making process (Macharis & Melo, 2010). Evaluating urban freight solutions in order to assess if they stand a chance, summons the need for a new approach taking into consideration different conflicting objectives from different stakeholders. The Multi-Actor Multi-Criteria Analysis (MAMCA) developed by Macharis (Macharis, 2000) combines the conventional multi criteria decision analysis (MCDA) with the notion of stakeholders in an explicit way. Within Straightsol (Strategies and measures for smarter urban freight solutions, EC FP7), a systematic impact assessment framework dedicated to freight transportation has been developed. The so-called City Distribution–Multi Actor Multi Criteria Analysis (CD-MAMCA) incorporates the city distribution actors and their objectives as the primary focus complemented with a MCDA. In view to compare the effectiveness of the different alternatives, the PROMETHEE GAIA Multi-criteria Analysis method was chosen. In this paper the specific framework dedicated to city distribution is explained with its applications to a case study within the Straightsol project. The relevant stakeholders within urban and interurban freight transport context are shown together with their important criteria. Besides the benefits obtained from the well-structured decision approach that offers the CD-MAMCA (Macharis et al, 2012) the results of the MCA reveal the preference order of the alternatives and allow developing specific recommendations for future freight measures. The same approach will be executed during the Straightsol project for the evaluation of seven demonstrations and aims to become an integrated framework for any evaluation. Keywords: CD-MAMCA, City Distribution, Multi-Actor Multi-Criteria Analysis, PROMETHEE. 91 TU4D (contributed) A Stochastic Choquet Integral Preference Model: SMAA-Choquet Corrente, Salvatore (*) Angilella, Silvia Greco, Salvatore University of Catania The Choquet integral is an aggregation function used in Multicriteria Decision Aiding (MCDA) to deal with interaction of criteria. The Stochastic Multiobjective Acceptability Analysis (SMAA) is an MCDA methodology used to deal with imprecision or lack of data in the decision problem at hand. In this paper we propose to integrate SMAA methodology with the Choquet integral preference model in order to obtain robust results according to available information. We show also how SMAA permits to induce not only a capacity representing importance and interaction between criteria, but also the common scale required by the Choquet integral. Keywords: Choquet Integral, Decision Making, Imprecise Information, Linear Programming, SMAA. 92 TU3B (invited) A Strategic Information System for Sustainable Supply Chain Taraszewski, Steve (*) Karpak, Birsen Youngstown State University Sustainable supply chains (SC) meet the needs of the present generation without compromising the ability of future generations to meet their own needs. There are three pillars of Sustainability: Economic, Environmental, and Social. The Economic pillar is the most common dimension addressed. Environmental life cycle impacts are increasingly being studied. However, research is still dominated by green/environmental issues and the social aspects and integration of the three dimensions are still rare. It is worth noting that no study has comprehensively addressed the three dimensions of sustainability (economy, society and environment), triple bottom line (TPL). There is an information flow component of supply chain management (SCM) systems. We contend that SCM systems are influenced by an information/knowledge sharing culture - the ability and willingness to share information and knowledge as to best-practices or potential opportunities or threats. Information systems (IS) ’touch’ a SC by facilitating the efficient and effective flow of data and information for elevating business processes supporting the entire SC. Also, IS offer the possibility of automating/informating/transforming the SC and its constituents. Few studies of sustainability incorporate the IS perspective. We propose a new strategic information system for sustainable supply chains. We have applied the proposed framework to an illustrative problem. We are evaluating three strategic partners with respect to TPL. Initial results illustrated that: Economic pillar is the most important followed by environmental and social performance dimensions; Partner C is the best strategic partner; innovation is the most influential performance factor followed by cost; among environmentally influential organizational practices recycle and reduce turned out the most influential; health care and labor equity are the most influential social factors, and cultural diversity is next. Keywords: Analytic Network Process, Strategic Information Systems, Supply Chain Management, Sustainability. 93 TH4B (contributed) A System Approach for Green Supplier Selection Based on the ANP Model with BOCR for the Electronics Industry in Taiwan Shih, Hsu-Shih (*) Chen, Chiau-Ching Lin, Yi-Chun Tamkang University This study deals with the problem of green supplier selection for the electronics industry in Taiwan. After defining a problem-solving procedure, we use an analytic network process (ANP) model with benefits, opportunities, costs, and risks (BOCR) merits to assess all factors for ranking suppliers. This study then illustrates a real-world case of supplier selection for the adapter of one type of notebook PC. We also set up a radar chart to show the priorities of the factors under the four merits so that managerial implications can be drawn. Firms in the past mostly relied on tangible criteria for making purchasing decision instead of considering the uncertain opportunities and risks in the evaluating process. The ANP model with BOCR merits can incorporate the uncertainties into the model. First, we compile 12 general criteria and 6 green criteria on supplier selection. Second, we take on the content validity ratio and factor analysis for acquiring elements and clusters under the BOCR aspects and then apply the ANP approach to process the independence and feedback among elements and clusters. Third, we execute a sensitive analysis. Lastly, we present the priorities of all factors under different merits. The model simultaneously considers the general criteria and green criteria in order to reward the environmental efforts of the suppliers. The results show that valued clusters are slightly different from the aspects of BOCR, i.e., “production adaptability”, “promise to customer”, “price” and “product reliability” are the major concerns, whereas green criteria are less valued. In addition, the ANP with BOCR can describe the criteria and factors in details and also consider the effects of uncertainties for a suitable decision. In the final part, we draw up suggestions for the stakeholders in the hope that they put forth more efforts on environmental protection in the future. Keywords: Analytic Network Process, Case Study, Environmental Decision Making, Sensitivity Analysis. 94 TH3F (contributed) A Tabu Search Algorithm for Multi Objective Open Vehicle Routing Problem Ozcetin, Erdener (*) Ozturk, Gurkan Anadolu University The Open Vehicle Routing Problem (OVRP) is a form of Capacitated Vehicle Routing Problem that turning depot from last node of route is not considered. The OVRP is NP-Hard and finding optimal solution with exact methods is extremely hard. In this study, we consider the OVRP as a multi-objective problem in which the objectives are minimizing total cost defined as a function of distance and maximizing the service level for customers. In order to scalarize the problem, conic scalarization which is powerful method to find pareto solutions is used instead of linear scalarization. For the real life case problem a variant of Tabu Search algorithm is developed. Obtained results show that proposed algorithm works efficiently in a short time period without ignoring service level. Keywords: Combinatorial Optimization, Tabu Search, Vehicle Routing Problem. 95 TH4B (invited) A Three-Stage Methodology to Analyze the Global Competitiveness of a Sector Ozaydin, Ozay (*) Cekyay, Bora Onsel Ekici, Sule Dogus University Ulengin, Fusun Kabak, Ozgur Istanbul Technical University This study focuses on developing a methodology to identify key success factors for improving the global competitiveness of a given sector, and to determine how much each critical factor should be ameliorated in order to achieve a specific competitiveness level with the minimum effort. In the first step, Analytical Hierarchy Process (AHP) is used to prioritize the most necessary sub-goals to maintain the competitiveness of the sector, which are determined by the main stakeholders of the sector. The second step attempts to quantify the sub-goals prioritized in the first step. This is achieved by the selection of the most relevant factors, included in the World Economic Forum (WEF) Competitiveness Index, for measuring the development of the corresponding sector in a given country. In the last step, a goal programming model is formulated to determine the minimum level of improvements in the selected WEF factors necessary to achieve a specific competitiveness level. A case study for Turkish Machinery Sector is also presented to illustrate the applicability of the proposed methodology. Keywords: Analytic Hierarchy Process, Competitiveness, Machinery Industry. 96 TH2B (contributed) A Two Objective Classification Approach based on Conic Functions Ozturk, Gurkan (*) Anadolu University Kasimbeyli, Refail Anadolu University Polyhedral conic functions were defined as a new class of functions for classification problems. The graph of a PCF is a cone and its level set at zero which is a convex polyhedron divides the whole space into two parts:inside and outside. A PCF is defined as a linear functions extended with l1 norm and characterised of its center point. If the center point is known in advance the construction of a function is straightforward, by solving a linear programming problem a PCF can be obtained easily. In this study a two phased approach is proposed. In the first phase according to different center points and norm term several conic functions are constructed then in the second phase a two-objective integer programming model is solved to find classification function. The performance of the proposed algorithm is shown on the literature test problems. Keywords: Classification, Data Mining, Mathematical programming. 97 WE1G (contributed) About an Implementation of a 3-Objective Linear Programming Solver Schenker, Sebastian (*) Zuse institute Berlin Despite the growing interest in multi-objective optimization and the wide availability of commercial and non-commercial single-objective solvers there exist only very few multiobjective (linear) programming solvers available to the scientific community. In this talk I want to present an implementation of a 3-objective linear programming solver that computes all extreme non-dominated points (and the corresponding weight space decomposition). In every iteration a weight is computed that leads to an unexplored extreme non-dominated point based on the approach of Benson and Sun. However, in contrast to the latter method we take the weight space decomposition into account in order to avoid a consecutively increasing linear system used for computing new weights. Instead, we look for new weights only on the weight space edges and keep the weight space decomposition updated after every iteration. The solver uses the lp interface of SCIP (framework developed at Zuse institute) that allows to connect different single-objective lp solvers (like soplex, cplex or gurobi) for the computation of the scalarized problem. The first part of the talk is concerned with a worst-case example based on deformed cubes that shows that we cannot expect a small number of extreme non-dominated points even for a fixed number of objectives. In the second part the basic ideas behind the solver and the solver itself are presented. Keywords: 3-Objective Linear Programming Solver, Linear Programming, Multi-Criteria Decision Analysis, Multi-Objective Optimization, Software. 98 MO2F (invited) Adaptive Guided Evolutionary Multi-Objective Optimization Siegmund, Florian (*) University of Sk¨ovde Deb, Kalyanmoy Michigan State University Ng, Amos University of Sk¨ovde In Multi-objective Optimization many solutions have to be evaluated in order to provide the decision maker with a diverse Pareto-front. In Simulation-based Optimization the number of optimization function evaluations is very limited. If preference information is available however, the available function evaluations can be used more effectively by guiding the optimization towards interesting, preferred regions. One such algorithm for guided search is the Reference-point guided NSGA-II. It takes reference points provided by the decision maker and guides the optimization towards areas of the Pareto-front close to the reference points. We propose several extensions of R-NSGA-II. In the beginning of the optimization runtime the population is spread-out in the objective space while towards the end of the runtime most solutions are close to reference points. The purpose of a large population is to avoid local optima and to explore the search space which is less important when the algorithm has converged to the reference points. Therefore, we reduce the population size towards the end of the runtime. R-NSGA-II controls the objective space diversity through the epsilon parameter. We reduce the diversity in the population as it approaches the reference points. In a previous study we showed that R-NSGA-II keeps a high diversity until late in the optimization run which is caused by the Pareto-fitness. This slows down the progress towards the reference points. We constrain the Pareto-fitness to force a faster convergence. For the same reason an approach is presented that delays the use of the Pareto-fitness: Initially, the fitness is based only on reference point distance and diversity. Later, when the population has converged towards the Pareto-front, Pareto-fitness is considered as primary-, and distance as secondary fitness. We combine the proposed approaches and evaluate them on benchmark functions and in a case study of an industrial production line for car manufacturing. Keywords: Adaptive Strategies, Evolutionary Multi-Objective Optimization, guided search, Reference Point, Simulation-Based Optimization. 99 MO5E (award) Advanced Metaheuristics for Multi-objective Optimization: Design, Analysis and Application Durillo, Juan J. (*) University of Innsbruck The optimization of problems involving two or more conflicting criteria consists of two phases: to compute the Pareto front-solving-, and to select a subset of points from it according to user preferences-decision making-. Computing the Pareto optimal front is impractical in most problems because of several reasons: it may consist of infinite points and it can be computationally expensive. On account of this, approximations to the optimal Pareto front are used. From a decision making point of view, the interest is to obtain approximations that are useful enough to choose satisfactory solutions. To this end, not all approximations are valid but those with a good convergence and diversity are seek. Convergence reflects optimality: the closer to the optimal Pareto front the better. Diversity has to do with the fact that the approximation consists of a finite number of points. The better these points cover the optimal Pareto front, and the better they are spread alongside it, the higher the success. An approximation with these two properties is then an important tool from which a decision maker may benefit, revealing points where the trade-off among the criteria is the most attractive. Therefore, the success of solving a given multi-objective optimization problem is strongly tied to the triumph of this phase. In addition, the time for solving problems still remains a wall to bring down specially for real-world problems that rely on expensive simulations and non-linear formulations. Researches are then challenged to design advanced methods able of computing these high quality Pareto fronts while reducing the required time to a solution. Two non-orthogonal ways to this end exist: reducing the number of required evaluations without damaging the quality of the computed results and applying parallelism. This talk aims to describe some advanced parallel metaheuristic methods targeted to face the aforementioned challenges and their application to solve real-world problems. Keywords: Applications, Metaheuristics, Multi-Objective Optimization, Parallelism. 100 MO4G (contributed) Adversarial Risk Analysis on Transport Infrastructures Cano, Javier (*) Rey Juan Carlos University R´ıos Insua, David Universidad Rey Juan Carlos Tedeschi, Alessandra DeepBlue Srl Turhan, Ugur Anadolu University In this work we analyze how the authorities of a small airport are concerned with terrorist threats against the airport ATC Tower, aimed at taking hold of air traffic controllers before or during flight control operations. Consequences of this could be severe: (1) crisis for air traffic operations in the air field and airspace; (2) flight safety negatively affected; (3) air traffics cancelled or diverted to other ATC unit or air fields, with enormous economic, social (in extreme cases, even in terms of human lives) and image costs. As a way to mitigate the impact of such menace, preventive measures, such as security checks and surveillance cameras, are being implemented. They have considerable associated costs but, by deploying them, authorities expect to deter the actions of the terrorists. The multiobjective nature of the problem becomes a more delicate issue as human lives and high costs are concerned. We model the problem as a particular case of a Sequential Defend-Attack model within the Adversarial Risk Analysis framework, in which authorities (the Defender) would deploy a portfolio of preventive measures. Then, the terrorists (an intelligent Attacker), having observed them, would follow a given terrorist strategy. We shall assume that the consequences for the Defender will depend on the effort in implementing their protective actions and on the mitigated result of the terrorist attacks. Similarly, the consequences for the Attacker will depend on the effort in developing their terrorist actions and their result. Both the Defender and the Attacker are regarded as expected utility maximizers. We present numerical results, giving advice for devising a multiobjective security contingency plan. The model is open to extensions, as e.g. larger installations, with more than one critical installations to be defended, new threats (more than one intelligent attacker), or additional countermeasures deployed by different agents (more than one defender). Keywords: Adversarial Risk Analysis, Airport Security, Cost Utility Analysis, Risk Analysis, Utility Theory. 101 TH2E (invited) Agent Assisted Interactive Multiobjective Optimization for Computationally Demanding Problems Ojalehto, Vesa (*) University of Jyv¨askyl¨a Miettinen, Kaisa University of Jyvaskyla Podkopaev, Dmitry University of Jyv¨askyl¨a A computationally expensive multiobjective optimization problem can be approached by replacing it with a computationally inexpensive surrogate problem, which approximates the set of Pareto optimal solutions of the original problem. This replacement enables applying interactive solution methods on account of faster interaction between the decision maker (DM) and the solution method. Decreasing computing time, however, is achieved at the expense of reducing the precision. Besides that, such an approach does not eliminate the cognitive difficulties experienced by the DM in problems with high dimensions of objective and design spaces. In order to overcome mentioned difficulties, we introduce multiple independent agents which provide the DM with information on the accuracy of solution candidates, learn DM’s preferences and based on that, improve the accuracy of the surrogate problem in those areas of the decision space which may be interesting for the DM. Furthermore, the agents can assist the DM by generating new Pareto optimal solution candidates based on the learned preferences. We demonstrate the agent assisted interactive decision making approach by applying the interactive NIMBUS method to an example problem of process engineering based on simulation of a generic two-stage separation process using the APROS dynamic plant-wide process simulation software. Keywords: Computational Cost, Interactive Methods, Multi-Agent Systems, NIMBUS, Preference Learning. 102 WE1C (contributed) An Algorithm to Determine Efficiency Scores in Large DEA Data Sets Yildiz, Gazi Bilal (*) Erciyes University Soylu, Banu Erciyes University When there are huge number of Decision Making Units (DMUs) and large number of input/output, identifying efficient units via a DEA model could be computationally complex. In this study, we propose a new algorithm to determine efficient DMUs and to estimate efficiency scores of inefficient units. Initially, we partition the original problem into sub-problems and solve these sub-problems to construct a set of initial efficient units. This set leads us to estimate the efficiency scores of other units. We apply the algorithm over large data sets and compare the performance with classic DEA and a decomposition algorithm from the literature. We present our results. Keywords: Data Envelopment Analysis, Multi-Criteria Decision Making. 103 TH4E (contributed) An Assessment of Multipurpose Water Resources Using Multicriteria Analysis Ribas, Jose Roberto (*) Pinheiro, Mariana Universidade Federal do Rio de Janeiro Severo, Juliana Ribas Delta Energia Water is essential for life; therefore, its management has attracted the attention of researchers worldwide. When considering large reservoirs formed by hydroelectric projects, the simultaneous use of water is a natural way to attend different needs of the local population, which, in some cases, are conflicting. The consideration of one specific use of water is a collective concern; consequently, its importance must be evaluated on a comparative basis, weighted by a set of relevant benefits provided to the ecosystem. This study describes the combination of two qualitative techniques (in-depth interview and content analysis) with the FAHP method, on attempting to rank a set of alternatives of a multipurpose water resource. A study was made with the reservoir of Corumba hydroelectric power plant, located in the Central region of Brazil, which borders eight municipalities. The multicriteria model was specified with the assistance of 43 experts classified among four groups (power plant owners; federal agencies experts; municipal mayors and officers; local population representatives). Initially, the qualitative assessment indicated eleven benefits required by the ecosystem and ten possible uses for the reservoir (electricity generation, water supply and irrigation were excluded), too complex to be treated by a multicriteria technique. The dimension of the model was reduced by means of a statistical cut off, resulting in six benefits and five uses. The metrics were done through a pairwise comparison using a nine-point Saaty’s scale. The FAHP technique estimated the weights, which resulted in four different ranks of benefits and uses. Such conflicts appointed some actions required to implement sustainable uses. The study was oriented to attend the local population demand and to protect the ecosystem as a whole. The qualitative data extracted from the interviews and the compliance with the Brazilian regulation were used as inputs to build up the recommendation. Keywords: FAHP, Multi-Purpose Water Resource, Sustainability Assessment. 104 TH2G (invited) An Efficient Box Algorithm for Discrete Tricriteria Optimization Problems D¨achert, Kerstin (*) University of Wuppertal Klamroth, Kathrin University of Wuppertal General multi-objective optimization problems are often solved by a sequence of parametric single objective problems, so-called scalarizations. If the set of nondominated points is finite, and if an appropriate scalarization is employed, the entire nondominated set can be generated. Ideally, the total number of scalarized subproblems depends polynomially on the number of nondominated points. In the bicriteria case approaches are known which, given an appropriate initial search space, require the solution of at most 2|N |−1 subproblems, where N denotes the nondominated set of the underlying problem. Thereby, |N | subproblems are solved to generate all points in N , and the additional |N |−1 subproblems are needed to ensure that no further nondominated points exist between the already generated ones (Ralphs et al., 2006). Up to now, no polynomial bounds are known for higher dimensional problems. The best known approach has a theoretical bound of (|N | + 1)m−1 subproblems for problems with m objectives (Laumanns et al., 2006). We present a new procedure for finding the entire nondominated set of tricriteria optimization problems for which the number of scalarized subproblems to be solved is bounded by 3|N | − 2. This is achieved by the definition of a new split criterion which allows to exclude redundant parts of the search space. It can then be shown that the number of boxes, into which the search space is decomposed, depends linearly on the number of nondominated points. Keywords: Multi-Objective Optimization. 105 MO2C (contributed) An Enumerative Cutting Plane Approach to Integer Linear Vector Optimization Problems Habenicht, Walter (*) University of Hohenheim The approach presented in this paper is based on the concept of intersection cuts, which is well known in classical integer linear programming. The method starts by generating an optimal basic solution of the lp- relaxation for some reasonable weighted sum of the objectives. The optimal basic solution is presented to the decision maker and he is asked to define a region around the optimal solution in objective space. In fact, we assume this region will be a hypercube. The feasible integer points in the interior of this hypercube are enumerated. The efficient ones are presented to the decision maker, and he is asked to choose the best one out of them. Then, the hypercube serves as a cut-generating set. The generated cut cuts off only those feasible integer points that have been enumerated before. Comparing the best feasible integer solution found so far with the (non-integer) vertices of the facet generated by the cut allows us to decide whether to stop the process or not. If the best feasible solution it preferred to the best vertex of the facet we can assume that under rather customary conditions the best solution is found. Otherwise, the best vertex of the facet is used as the next basic solution of the lp-relaxation and a next iteration is performed. Keywords: Integer Programming, Interactive Methods, Mixed Integer Programming, Multi-Criteria Decision Making, Multi-Objective Optimization. 106 TU2C (contributed) An Evaluation of the Situation Judgment Test as a Multidimensional Measure of Decision Making Reinerman-Jones, Lauren (*) University of Central Florida Parchment, Avonie University of Central Florida Teo, Grace University of Central Florida Decision making is a skill often sought to quantify. This goal is achieved primarily through the use of tasks or questionnaires. The challenge with either of these approaches is that the metrics are often unidimensional. A widely-used multidimensional measure is the Situation Judgment Test (SJT) which comprise work scenarios, each with various possible response options. These scenarios entail complex, real-world problems and respondents must decide on the best choice for each. The scenarios and response options are modifiable by testers, and the problem with this is that the scenarios may not be content valid. Although the criterion validity may still indicate the usefulness of the SJT, performance data needed to determine validity may not be available till much later. Also, such post-hoc analysis is biased as it requires the tool to be used before its evaluation, and personnel may have self-selected out or be released of duties before the validation completes. This may result in the wrong test being used, which is very expensive and time consuming. The aim of the present study is to explore a methodology for evaluating the content and construct validity of SJT scenarios that could later be used to assist in the selection of good decision makers. This was accomplished by administering six SJT scenarios to 90 participants, along with the Adult Decision Making Competence (ADMC; Bruine de Bruin, Parker & Fischhoff, 2007), Cognitive Reflection Task (CRT; Frederick, 2005), Decision Outcome Inventory (DOI; Bruine de Bruin, Parker & Fischhoff, 2007) and moral judgment items (Greene, Morelli, Lowenberg, Nystrom & Cohen, 2008; Starcke, Polzer, Wolf & Brand, 2011). Performance on those six scenarios correlated with two of the scales in the ADMC, the CRT, and the moral judgment items. The results indicated that these measures can help shed light on the content validity of the SJT. Future research should examine other measures for evaluating SJTs prior to implementation. Keywords: Decision Making, Measurement, Multi-Dimensional, Psychology, Situational Judgment. 107 WE2G (invited) An Evaluation of Website Upgrade Options. A Case Study Comparison of ANFIS and RIMER Xu, Ling (*) The University of Manchester Chen, Yu-Wang The University of Manchester Sabin, Andrada The University of University In this paper, two methods, Adaptive Neuro Fuzzy Inference System (ANFIS) and Belief Rule-Base Inference Methodology Using the Evidential Reasoning Approach (RIMER), are investigated in a real case scenario of supporting a strategic decision process: Sapico Chemical International’s website upgrade. Literature is first analysed to structure the problem and a bilateral approach is employed taking into account both company and its customer opinions. The established framework is then further refined following the opinions and requirements of Sapico. The refined framework is then populated with the data obtained from Sapico’s manager, experts and the aggregated results of surveys conducted by Sapico. Two comparative investigations are carried out. In one investigation, options for upgrading the company web site are evaluated through two criteria: customer and company point of views. In the other, the options are evaluated using the same two criteria except that customer point of view is further decomposed into sub-criteria. We name the resulted models 1-level and 2-level models respectively. In both investigations, instead of aggregating criteria, we use fuzzy rules for ANFIS and belief rules for RIMER and their respective inference methods for the evaluation. The overall results show that 1-level ANFIS and RIMER models have very similar performances. The 2-level ANFIS model performs slightly worse than the other three. When forecasting aggregated customer opinions, the 2-level RIMER model performs considerably better than the 2-level ANFIS. With regard to explanatory power, RIMER models perform better than ANFIS ones. RIMER is able to explain the inherent risks and advantages within every output score. The same high explanatory power can also be noticed in RIMER when looking at the trained rule bases. We have explored other problem contexts that the solution can be applied. In addition to this, we have also provided Sapico with a ranking of the options for an upgrade. Keywords: Aggregation Schemes, Artificial Intelligence, Belief Rule-Based methodology, Fuzzy Reasoning, Neural Networks, Structuring Decisions. 108 FR2F (invited) An Evolutionary Multiobjective Optimization Algorithm for Diverse Solutions Chmielewski, Hana (*) North Carolina State University Piper, Brian E.B. North Carolina State University Ranjithan, S. Ranji North Carolina State University Issues with multiple objectives and incompletely defined problems are frequently encountered when solving engineering problems. Evolutionary multiobjective optimization (EMO) algorithms provide a possible method for addressing these difficulties. In this paper, EMODS (Evolutionary Multiobjective Optimization Algorithm for Diverse Solutions) is proposed for the purposes of approximating a Pareto front and locating nearlynondominated alternative solutions with diverse solution characteristics. A hypervolumebased fitness calculation method is proposed that assigns fitness to all solutions in a population. This fitness method allows solutions with similar levels of performance to compete for selection based on decision space diversity. An archive is described to retain nondominated solutions and nearly-nondominated alternative solutions. Measures of decision space diversity are defined to capture the degree to which solutions are spread through the decision space and the distinctness of the alternative solutions that are located. The algorithm is evaluated on a range of test problems and compared against other diversity enhancing EMO algorithms in the literature. The results show that the proposed algorithm is competitive or superior to the compared EMO algorithms in terms of discovering a high-quality nondominated set. Additionally, results of the decision space diversity metrics show that the proposed algorithm achieves greater solution diversity than other methods and discovers alternative solutions that are distinctly different from other solutions in the decision space. Keywords: Evolutionary Computations, Generating Alternatives, Genetic Algorithms, Multi-Objective Optimization, Solution Diversity. 109 WE2C (contributed) An Extension of the Fuzzy DEMATEL Method for Group Decision Making Almulhim, Tarifa (*) The University of Manchester Mikhailov, Ludmil The University of Manchester Xu, Ling The University of Manchester The Fuzzy Decision Making Trial and Evaluation Laboratory (Fuzzy DEMATEL) method is one of the Multi Criteria Decision Making tools, which can structure complex problems through visual structural model and analyze the causal relationship between factors. In the existing literature, the Fuzzy DEMATEL method assumes that all the decision makers are equal in their relative weights of importance in the problem considered. However, in some complex group decision making problems this has its drawbacks. This paper proposes a fuzzy DEMATEL method for group decision-making with non-equal importance weights of individual members in the group. The directed influential degrees between pair-wise factors are expressed as fuzzy numbers in order to model the uncertainty and imprecision judgments from the group’s members. Unlike the known Fuzzy DEMATEL, the proposed method considers the importance weights of individual members in the group and presents those weights as either crisp numbers or fuzzy numbers, which are used in a fuzzy aggregation procedure. In this procedure, the Weighted Arithmetic Mean method is applied for individual members’ weights of importance, represented as crisp values, whereas, the Fuzzy Weighted Arithmetic Mean method is used for the fuzzy weights. For assisting decision makers to apply the proposed method, a prototype of a decision support tool is developed using the MATLAB Software. A comparison between the proposed and the traditional Fuzzy DEMATEL methods is illustrated with a numerical example. The results show that the proposed method is more suitable and reasonable than the traditional Fuzzy DEMATEL method. Keywords: Fuzzy DEMATEL, Group Decisions Making, Importance Weights, MultiCriteria Decision Making. 110 TU3F (contributed) An Improved MOPSO Algorithm for Multiobjective Bilevel Linear Problems Alves, Maria Jo˜ao (*) Univeristy of Coimbra Bilevel programs allow to model optimization problems with a hierarchical structure between two decision makers (the leader and the follower), who make decisions sequentially. The two decision makers pursue different objectives in a non-cooperative manner by controlling different sets of variables, subject to interdependent constraints. The leader makes his/her decisions first, thus reducing the set of feasible choices available to the follower. The follower reacts by choosing an optimal candidate according to his/her objective function on the feasible region restricted by the leader. This choice also affects the value of the leader’s objective function, so the leader must anticipate the reactions of the follower. A bilevel programming problem is very difficult to solve, even the linear case. Multiple objectives at one or both levels add further difficulties to deal with the problem. In this work we address the bilevel linear programming problem with multiple objective functions at the upper level and a single objective function at the lower level. We have previously developed and tested a multiobjective particle swarm optimization (MOPSO) algorithm to solve this type of problems, which aims to generate a good approximation of the entire Pareto front. A new technique for the “global best” selection within the MOPSO algorithm was tested, which revealed a good convergence towards the Pareto front but not so good in the diversity of the obtained solutions. We now propose an improved MOPSO algorithm for the same class of problems, which uses a hybrid strategy for the “global best” selection and an adaptive mutation operator. This operator uses information from the candidate solutions in the “nondominated archive” and also properties of the problem to adjust probabilities of mutation for the values of variables. The incorporation of these mechanisms improves the previous algorithm on all tested problems. The algorithm and computational results are presented. Keywords: Hierarchical Optimization, Metaheuristics, Multi-Objective Optimization. 111 WE1B (contributed) An Integrated AHP&VIKOR Method for Hospital Site Selection Yılmaz, Hafize (*) Hali¸c University G¨ ul, Sait Halic University Topcu, Ilker Istanbul Technical University Optimal site selection issue is one of the most important decision making problems for the investors because of the problem’s complicated decision processes relating with the number of criteria which should be considered attentively while making final decision. Particularly, the site selection for a new hospital has many criteria that are required to be considered, i.e. its investment cost, impact to the nature. In case of public investors like general or local government and municipal authority, erroneous or inefficient decisions may cause excessive costs on city budget or damages to the environment. In this study, we propose an integrated multiple attribute decision making (MADM) method under fuzzy environment. Today’s world encounters more complex problems with compared to the past. In the same manner, the decision problems have some difficulties to be defined precisely, because they can be affected from different and discrete factors or conflicting criteria and have multiple objectives. In this study, we utilize fuzzy logic in order to deal with Analytic Hierarchy Process (AHP) and Vise Kriterijumska Optimizajica I Kompromisno Resenje (VIKOR) methods under imprecise environment. Also, we handled the hospital site selection problem as a group decision making problem and appealed more than one decision experts’ knowledge and experiences. The usage of MADM methods for hospital site selection is a nadir research area in literature. This study proposes a novel approach for hospital site selection by utilizing fuzzified MADM methods in an integrated way under imprecise environment. Briefly, AHP is used to determine the importance weights of the selection criteria affecting the site of the hospital; and VIKOR analyses the alternative sites and suggests the closest one(s) to the ideal solution. Keywords: Analytic Hierarchy Process, Fuzzy Logic, Hospital Site Selection, VIKOR. 112 MO4B (invited) An Interactive Algorithm for Bi-objective Routing Problems Tezcaner, Diclehan (*) Middle East Technical University Koksalan, Murat Middle East Technical University In this study, we develop a branch and bound based algorithm for bi-criteria integer programs that finds the most preferred solution of a decision maker (DM) with a quasiconvex preference function to be minimized. We demonstrate our algorithm on the bi-objective routing problem for unmanned air vehicles (UAV) which is assumed to travel on a terrain discretely approximated by grids. The algorithm finds the best route of a DM between a number of targets in the presence of two objectives. Finding the efficient paths between targets can be modeled as a multi-objective shortest path problem. Finding the efficient tours made up of the efficient paths can be modeled as a multi-objective traveling salesperson problem. Therefore, multi-objective routing problem is a combination of these two combinatorial problems. In our interactive algorithm, we need all the efficient shortest paths between all pairs of targets as an input to the branch and bound procedure we develop. We establish some rules that decrease the number of efficient paths required for the solution of the branch and bound algorithm. We also address the more general case where the terrain is defined in a continuous space for the UAV routing problem. We study the problem of identifying efficient paths for this case. Keywords: Bi-Objective Routing, Combinatorial Optimization, Interactive Methods. 113 MO2C (contributed) An Interactive Algorithm for Multi-objective Integer Programs Koksalan, Murat (*) Middle East Technical University Korhonen, Pekka Lokman, Banu Wallenius, Jyrki Aalto University We develop an interactive algorithm to find the most preferred solution for multi-objective integer programs. We assume that the decision maker’s (DM’s) preferences are consistent with a nondecreasing quasiconcave value function. Based on the properties of the value function and pairwise preference information obtained from the DM, we generate constraints to eliminate inferior regions with respect to cones. We also develop a variation where we eliminate regions that are close to being dominated by the cones in addition to the regions dominated by the cones. The algorithm continues iteratively and approximates the most preferred solution for any multi-objective integer program with a desired level of accuracy provided that the DM’s preferences are consistent with a nondecreasing quasiconcave value function. The algorithm allows the DM to change the desired level of accuracy throughout the solution process. We test the performance of the algorithm on a set of multi-objective combinatorial optimization problems. It performs well in terms of the quality of the solution found, the solution time, and the required preference information from the DM. Keywords: Integer Programming, Interactive Methods, Multi-Objective Optimization. 114 MO2C (contributed) An Interactive Algorithm for Multiobjective Mixed Integer Programming Problems Ozpeynirci, Ozgur (*) Izmir University Koksalan, Murat Middle East Technical University Lokman, Banu Aalto University In this study, we aim to find the most preferred nondominated point of a multiobjective mixed integer programming problem with p objective functions. Without loss of generality, we assume all objectives are to be maximized. We first consider an underlying linear value function. It is known that an extreme supported nondominated (ESN) point that is preferred to its adjacent ESN points is the most preferred point for an underlying linear value function. We first develop an exact procedure for finding all adjacent ESN points of a given ESN point. Utilizing this procedure, we develop an interactive algorithm that operates in the objective space and starts from an arbitrary extreme supported nondominated (ESN) point. It evaluates adjacent ESN points using past and current preference information gathered through the decision maker interactions. It searches for preferred regions utilizing all the available information and terminates when it finds the ESN point preferred to all its adjacent ESN points. We also generalize the algorithm for the quasiconcave value function case. In this case, we again identify an ESN point that is preferred to all its adjacent ESN points. We then search for preferred unsupported nondominated points within the neighborhood of the identified ESN point. Interacting with the decision maker, the algorithm converges to the most preferred point. We conduct computational tests on multiobjective assignment and knapsack problems. Keywords: Adjacency, Interactive Methods, Value Function. 115 FR3C (contributed) An Interactive Approach to the Bi-objective Inventory Routing Problem Huber, Sandra (*) Helmut-Schmidt-University Geiger, Martin Josef Helmut-Schmidt-Universtiy Hamburg Sevaux, Marc Universit´e de Bretagne Sud The talk presents an interactive approach for the bi-objective inventory routing problem (IRP). The IRP combines delivery quantity decisions with a vehicle routing problem into a simultaneous model. In this problem setting, a considerable tradeoff between the minimization of the inventory levels on the one hand, and the minimization of the transportation effort (i.e. the traveled distances) on the other hand is present. A local search approach on the basis of a multi-point hillclimber is used for computing solutions. After each solution construction phase, the decision maker is actively involved in the intermediate decision making process. He/she is allowed to state personal preferences, thus guiding the subsequently derived search direction. We investigate the effectiveness of a reference point-based approach for the problem at hand. A prototypical implementation of the concept into a running system has been made available, and experiments can and have been conducted on its basis. In addition to the study of the above mentioned logistical problem, we also look into ideas of how to simulate a decision maker in Silico. This related to our motivation of obtaining more general insights into the behavior of interactive search-and-decision-making systems for such multi-objective combinatorial optimization problems. Keywords: Inventory Routing Problem, Logistics, Multi-Objective Optimization, Transportation. 116 TU3G (contributed) An Interactive Tool for Multi-Objective Linear Programming Henggeler Antunes, Carlos (*) University of Coimbra Alves, Maria Jo˜ao Univeristy of Coimbra In addition to exposing students to concepts and methods, teaching optimization algorithms in engineering, economics and management courses requires a hands-on approach using tools that enable them experimenting and exercising critical thinking. This is quite important in teaching multi-objective optimization since the non-dominated solution set needs to be identified, desirably in a constructive manner that may provide further information about the trade-offs involved. In practice, presenting a decision maker (DM) with a large set of non-dominated solutions does not convey usable information for actual decision support. The involvement of the DM by providing indications about his/her preferences is essential to guide and reduce the scope of the search, thus minimizing both the computational effort and guaranteeing that the new solutions computed are more in agreement with his/her (evolving) preferences. By intertwining computation and judgment phases, interactive methods allow for a progressive shaping of the DM’s preferences as the selective characterization of non-dominated solutions develops. Therefore, interactive computer tools are indispensable to enhance the students’ technical skills and provide training as DMs, integrating different interactive methods to deal with MOLP using different solution computation techniques, search strategies, preference elicitation requirements, visual interaction mechanisms and result displays. This is being used as the entrance door for improving students’ understanding of the main issues at stake in MOP, before progressing to more technically demanding topics, in which the main theoretical concepts can be perceived through experimentation at their own pace. Moreover, it allows students playing the role of actual DMs and exploit different search strategies and trade-offs leading to distinct non-dominated solutions. A computer tool is presented integrating distinct interactive approaches in a consistent operational framework. Keywords: Decision Support Systems, Education, Interactive Methods, Multi-Objective Optimization. 117 TH2D (invited) An Optimization Model for Group Decision Making: Weighting and Ranking with Incomplete Fuzzy Preference Relations Ruiz-Tagle, Mauricio (*) Universidad Austral de Chile Dopazo, Esther Universidad Polit´ecnica de Madrid We consider a finite set of alternatives that have to be ranked according fuzzy preference relations given by multiple experts. In real-world problems it is usual that experts are not able to efficiently provide their preferences on some aspects of the problem because large number of alternatives, limited expertise related to some problem domain, not available data, etc. In this paper a model is presented to generate a group priority vector of the alternatives that reflects as best as possible the incomplete fuzzy preference relations given by the experts. The resulting optimization problems are solved by using goal programming techniques. Furthermore we discuss sufficient conditions on the structure of the data available that guarantee the existence of a priority vector. Keywords: Fuzzy Sets, Goal Programming, Group Decisions Making, Meta-Goal, Preference Aggregation. 118 WE1D (invited) Analytical Approximation of the Influence of Capacity on the Economic and Ecological Performance of Bioenergy Plants Lauven, Lars-Peter (*) University of G¨ottingen In this work, economically optimal capacities for bioenergy plants are determined by applying a nonlinear optimization approach. Both economies of scale and required transportation distances are approximated with appropriate functions to identify optimal capacities with regard to the required investment, input and output prices and the available amount of biomass in the proximity of the envisioned plant. While larger bioenergy plant capacities lead to improved specific production costs, they also increase biomass transportation distances. In addition to identifying economically optimal production capacities for a number of scenarios, the investigation therefore also aims to quantify the effect on transportation-related GHG emissions and discusses side-effects of concentrated energy plant production. Keywords: Bioenergy, Epsilon-Constraint Method, Greenhousegas Emissions, Non-Linear Programming, Plant Capacity Planning. 119 TH4B (contributed) Analyzing the Provision of Agricultural Public Goods: The Case of Irrigated Olive Groves in Southern Spain Villanueva-Rodr´ıguez, Anastasio (*) Inst. Agric. and Fisheries Research Arriaza, Manuel Institute of Agricultural and Fisheries Research G´omez-Lim´on, Jos´e A. University of Cordoba Nekhay, Olexandr University of Cordoba The analysis of the joint production of private and public goods from farming activities is a fertile research field in agricultural and environmental economics aiming at supporting public decision-making regarding a better governance of the agricultural sector. These joint production processes are typically characterized by a high level of complexity derived from the intense relationship between the productions of both kinds of output. Hence, an integrated approach is strongly recommended to study the provision of agricultural public goods and the design of public intervention in this sector. Among the methodologies that apply such approach in this research field, the Analytic Network Process (ANP) is one of the most promising because it allows to consider interdependencies among different elements of the system studied. Here, we use the ANP to assess how the public goods provided by farming activities can be affected by farmers’ decision-making, and thus how effective policy intervention could be. This would be helpful for policy-makers to identify intervention priorities, enabling policy instruments to be primarily focused on the more modifiable public goods provisions. This methodological approach is empirically applied to a particular farming system; the irrigated olive groves of southern Spain. Through the ANP network designed for this case study, we have identified the public goods more influenced by farmers’ decisions, namely soil fertility, farmland biodiversity and visual quality of agricultural landscapes. In addition, the most influential factors affecting public goods provision have been identified. Particularly, agricultural practices dealing with fertilization, soil and irrigation management, as well as structural variables regarding farm size, density of planting and variety of olive tree are worth acknowledging. These results are useful for supporting agricultural policy decision-making to enhance an adequate management of this farming system. Keywords: Agricultural Public Goods, Analytic Network Process, Irrigated Olive Grove , Southern Spain. 120 MO4E (contributed) Application of an Intelligent Decision System to Assess Students’ Academic Performance in University Early Admission Chen, Yu-Wang (*) The University of Manchester Chen, Yue Shanghai Jiao Tong University Xu, Ling The University of Manchester Yang, Jian-Bo The University of Manchester This paper presents how the evidential reasoning approach of multi-criteria decision analysis, with the support of the Intelligent Decision System (IDS), can be used to support the data analysis and decision making process in university early admission. Following a literature review on the academic performance assessment problem and relevant statistical and artificial intelligence techniques, the paper introduces the criteria which are used to assess students’ academic performance in practice, and then describes the process of using IDS for model implementation, information collection and aggregation, and graphical presentation of outcomes. The academic performance assessment with the application of IDS can generate informative results to support the decision making process in practical university early admission. Keywords: Academic Performance Assessment, Early Admission, Evidential Reasoning, Intelligent Decision System. 121 TH3E (contributed) Applying of Fuzzy Logic to Decision-Making Control of the Ship Motion Malecki, Jozef (*) Polish Naval Academy Fuzzy logic systems find wide practical applications to making decision ranging from soft control in consumer products to accurate control and modeling of complex nonlinear systems of ship. A paper showed a possibility of using a fuzzy logic environment as a method making-decision to solve the problem related to determining an optimal trajectory in ship motion. It is possible to effectively solve tasks of determining a trajectory movement of a ship as a multistage decision-making process in a fuzzy environment. A paper explores fuzzy logic and how it helps navigators to solve control problems commonly found in autopilot applications – automatic control systems of the ship. Fuzzy logic, which mathematically emulates human reasoning, provides an intuitive way to design function blocks for intelligent control systems of the ship motion. Automatic control systems deploying fuzzy logic can improve the management of uncertain variables, such as making decision to control of the ship. Basic modules of the automatic control system consists of a Fuzzy Autopilot, a Thrust Distribution Module, a Propulsion System, the Ship and Filtering-Estimation Module. The Autopilot computes command signals τ d comparing desired ship’s position xd, orientation and velocities with their current estimates. Corresponding values of propellers thrust f are calculated in the Thrust Distribution Module and transmitted as control input to the Propulsion System. The propulsion system generates a control signal consists of thrust forces actuators of the Ship. On the ship are acting disturbances d. The control autopilot adjustments are modified by a position and velocity data from the Filtering-Estimation Module. The paper consists of the following sections. It starts with a brief description of dynamical and kinematical equations of the ship’s motion. Then a fuzzy control law and a power distribution algorithm are presented. Next some computer simulation results and conclusions are provided. Keywords: Applications of MCDM, Control, Decision Making, Dynamic Systems, Fuzzy Logic, Marine Engineering Systems. 122 FR3E (contributed) Applying the Utadis Method to the Identification of Key Factors of Organisational Commitment: A Case Study in Petrobras Autran M. Gomes, Luiz F. (*) Ibmec/RJ Gon¸calves, Marcos E.L. Petrobras Rangel, Luis A.D. Fluminense Federal University The key objective of this study was to perform a classification of the items of an organisational ambience survey at Petrobras in relation to its impact on organizational commitment. The method used for sorting the items was UTADIS, using criteria defined from organisational commitment measurement questionnaires. The items with highest impact on organizational commitment were then sorted. The classification proposed was used as a basis for a discussion between employees and the HR team. That discussion led to defining the priorities for the ambience improvement action plan. The secondary objective of the study was creating an evaluation and prioritisation process which could be used in other management areas of the company. The overall results showed the applicability of the method to the problem, opening the possibility of it being used in other management areas of Petrobras. Keywords: Human Resources, Linear Programming, Multi-Criteria Sorting, UTADIS. 123 MO4D (invited) Approximations for the Multi-objective Spanning Tree Problem Ismaili, Anisse (*) LIP6 - UPMC The multi-objective spanning tree problem (MOST) is a very challenging problem due to the combinatorial blow-up of the Pareto set as the size of the problem increases. In the bi-objective case, exact non-polynomial two-phase methods hold a good time cost see e.g [Spanjaard et al. 2008]. These methods, first compute the set of Pareto optimal solutions which are supported by a linear combination of the multi-objective cost vectors, then apply branch and bound to compute the non-supported Pareto optimal solutions. Nevertheless, generalizing such 2-phases methods to problems involving p > 2 criteria is far from trivial and alternative solutions methods are worth investigating. Moreover, it is often useless to generate the whole Pareto set (that grows exponentially with the problem) and a more reasonable objective is to only find a small epsilon-approximated covering of the Pareto-set [Papadimitriou et al. 2000]. To the best of our knowledge, for p > 2, the best known pseudo polynomial algorithm addressing the MOST [Papadimitriou et al. 2000] relies on the matrix-tree theorem [Kirchhoff]. It computes the determinant of a Laplacian matrix over polynomials by using a dynamic programming algorithm without divisions [Mahajan et al. 1999, Rote 2001]. By rounding the costs, the resulting theoretical FPTAS is very expensive, despite polynomial time. After a simplification of this FPTAS, we experiment different dynamic programming schemes to compute either the exact Pareto-set or an epsilon-covering. Their optimal sub-structures are defined by the set of vertices which are covered and connected. The first scheme iteratively adds vertices one by one to the sub-structures. The second scheme merges the sub-structures. Both algorithms can be modified to build an epsilon-approximated covering of the Pareto-set. Ultimately, we experiment the contribution of multi-objective heuristic search to state spaces with similar definitions. Keywords: Approximation, Combinatorial Optimization, Multi-Objective Optimization, Spanning Tree. 124 MO4C (invited) Approximations of Set-Valued Optimization Problems L´opez Montoya, Rub´en (*) Universidad Cat´olica de la Ssma Concepci´on Hern´andez, Elvira Universidad Nacional de Educaci´on a Distancia (UNED) The aim of this work is to introduce a notion of convergence for set-valued mappings and to show that this notion is suitable for studying set-valued optimization problems via approximation procedures. Keywords: Efficiency, Mathematical programming, Pareto Optimal Solutions, Sensitivity Analysis. 125 TU4G (invited) Assessing the Efficiency of Equity Mutual Funds. A Fuzzy DEA Approach Baeza-Sampere, Ismael (*) Coll-Serrano, Vicente Universidad de Valencia M’Zali, Bouchra UQAM M´endez Rodr´ıguez, Paz University of Oviedo Although several DEA models have been proposed in the literature for mutual funds performance evaluation, few of them incorporate non-financial criteria. In this paper a fuzzy DEA model is used, allowing mutual funds relative performance evaluation in a more realistic and flexible way. We examine efficiency of 40 US large cap equity mutual funds based not only on financial variables but also on non-financial ones. To achieve this aim, we extend Basso and Funari’s mutual funds’ ethical level proposing a more reliable fuzzy measure of the Social, Environmental Responsibility degree of equity mutual funds. It relies on the Corporate Social Performance of the companies invested in by the mutual funds and on the Quality of the Management in terms of the transparency and credibility degree of the non-financial information provided by the mutual funds. We can conclude that Socially Responsible Mutual Funds show a better behaviour in terms of efficiency than conventional funds. Keywords: Efficiency, Equity Mutual Funds, Fuzzy DEA, Socially Responsible Investment. 126 TU5B (contributed) Assessment of Renewable Energies based on Fuzzy and Qualitative Multi-Criteria Decision Making: a Comparison of two Wethodologies Afsordegan, Arayeh (*) Cremades, L´azaro V. Sanchez, Monica Zahedi, Siamak Universitat Polit`ecnica de Catalunya Agell, N´ uria ESADE, Universitat Ramon Llull Renewable energies have become a crucial issue in energy sector research due to their increasingly positive impact on sustainability. Current studies focus on the impact of existing sources of renewable energies, as well as their comparison and possible combinations. The assessment and selection of the most suitable types of energy in a geographical area is a complex problem, involving technical, economic, environmental, political, and social criteria. To establish a methodology to select the best option depending on different criteria in each scenario, this paper compares two multi-criteria decision making (MCDM) approaches based on linguistic label assessment for renewable energies. The first approach consists of a modified fuzzy Topsis methodology introduced by Kaya and Kahraman in 2011. The second approach, introduced by Agell et al. in 2012, is based on qualitative reasoning techniques for ranking multi-attribute alternatives in group decisionmaking with linguistic labels. This second methodology has achieved interesting results in various managerial applications, but has not yet been applied in the energy sector. An example, based on data provided in a paper by Kaya and Kahraman in 2011, is used to illustrate the mechanisms employed in both approaches and analyse their similarities and differences. We consider seven energy alternatives, six renewable (solar, wind, nuclear, biomass, hydraulic, and combined heat and power) and one conventional. Both MCDM methods are performed on the basis of nine criteria: efficiency, energy efficiency, investment cost, operation and maintenance cost, NOX emission, CO2 emission, land use, social acceptability, and job creation – as conveniently weighted by a group of three experts using the analytic hierarchy process (AHP). The obtained results are presented and analysed using both methods in five scenarios. Keywords: Analytic Hierarchy Process, Energy System, Fuzzy Sets, Group Decisions Making, Multi-Criteria Decision Making, Qualitative Reasoning. 127 WE1B (contributed) Assessment of the Relationship between Planned Obsolescence and Product Eco-Design based on MCDA Casas P´aez, Pedro Nicol´as (*) Universidad Nacional de Colombia Cortes Aldana, Felix Antonio Universidad Nacional de Colombia Purpose: To evaluate the linkage between eco-design and desired obsolescence in electronic manufacturers and systems integrators. Design/methodology/approach: The Analytic Hierarchical Process is a Multi-Criteria Decision Analysis methodology, used to formalize the intuitive understanding of complex problems by the construction of a Hierarchical Structure. The decision model is presented, taking into account the stakeholders of a local electronic system integrator. Findings: The study develops a framework for measuring the relative importance of obsolescence models based on the application of eco-design concepts. Due the particular context and considerations presented on the study, the model may need to be customized to fit the organization’s structure. Research limitations/implications: The model presented is dependent on the concept of the decision-maker team, and the generalizability of findings on this model to other products or different environmental situations may be limited. Adding further characteristics related to the total product may change the presented perceptual weightings. Practical implications: This paper addresses one of the challenges of the green marketing and the product administration, in determining which obsolescence model is preferred when taking into account the environmental eco-design principles of the base product. Originality/value: This paper offers practical help by providing a decision model that assists in determining which obsolescence model is most preferred based on firm’s judgment around eco-design attributes and system design, as well as the prioritization of those attributes. Keywords: Analytic Hierarchy Process, Decision Making, Multi-Criteria Decision Analysis, Obsolescence, Product Development. 128 MO2E (contributed) Augmented Goal Programming with Multi Fuzzy Targets and Coefficients Mouslim, Hocine (*) Tlemcen University Belmokaddem, Mustapha Tlemcen University Melloul, Sakina Tlemcen University Decision making is becoming more and more complex for managers, mainly in the multicriteria problems. Multi-choice goal programming (MCGP), and Multi-segment goal programming (MSGP) are considered as powerful tools to solve this type of problems. In these models, the decision maker considers the multi targets and the coefficients (parameters) of each goal as precise and well known. However, in some of the decision making problems, there may exist situations where the manager could not be interested in presenting his prefered targets and coefficients in a precise manner. To deal with such situations, Tabrizi et al. (2012) have lately reformulated (MCGP) in order to model the targets fuzziness. In their model, the only functions which are used are the triangular membership ones which do not reflect adequately the decision maker’s (DM’s) preferences that are considered an essential element for modelling the parameters (targets and coefficients) fuzziness. In this paper, new ideas are presented to reformulate (MCGP), and (MSGP) in a single formulation in which the decision maker’s (DM’s) preferences are taken into consideration for modelling the targets and also coefficients for all the types of goals in an imprecise environment. Moreover, the concept of indifference and veto thresholds is used in the new formulation for characterizing the imprecision and the preferences of the goals. This model is called Augmented Goal Programming with multi fuzzy targets and coefficients (AGP). A numerical example is given to demonstrate the validity and strength of the new model. Keywords: Augmented Goal Programming, Decision Maker’s Preferences, Decision Making, Indifference Thresholds, Multi-Choice Goal Programming, Multi-Parameter Functions. 129 TU4F (contributed) Automating Mechanism Design using Evolutionary Multi-Criteria Optimisation Chandra, Arjun (*) University of Oslo Allmendinger, Richard University College London Lewis, Peter University of Birmingham Market-based interaction mechanisms realise the allocation of contended resources across multiple strategic agents operating in multi-agent systems or electronic commerce environments. Designing these mechanisms has traditionally been carried out by hand, and more recently by automation. Measuring the performance of these mechanisms requires them to be evaluated with respect to multiple conflicting system-wide criteria, which can often be non-linear, noisy, and expensive to compute. Automated mechanism design involves discovering mechanisms that perform well on such criteria, subject to constraints that limit the search space to a class of mechanisms of interest. We focus on mechanisms that induce the allocation of m identical resource units across n agents (m < n) in a unit-demand setting. Considered is the Vickrey-Clarke-Groves redistribution mechanism class, which is further constrained to being strategy-proof, non-deficit, and individually rational. The system-wide performance criteria are defined as the average case welfare, envy, and disproportionality rates exhibited by the mechanism, with uncertainty in the type profile of agents imposing noisy and expensive evaluations. Whilst previous research has looked at automated design using a single aggregated criterion, we propose the application of evolutionary multi-criteria optimisation, and exemplify this approach by providing approximations to Pareto-optimal redistribution mechanisms for the resource allocation scenario at hand. We show that the approach gives not only insights into the range of different redistribution mechanisms, but also reveals trade-off mechanisms with superior envy and disproportionality rates when compared with state-of-the-art redistribution mechanism designed both by hand and by previous automated mechanism design approaches. Keywords: Automated Mechanism Design, E-Commerce, Economics, Expensive Evaluations, Multi-Agent Systems, Multi-Objective Optimization, Redistribution Mechanisms. 130 TU4C (contributed) Best Practices for Spatial Multicriteria Evaluation: a Critical Review Ferretti, Valentina (*) Politecnico di Torino Montibeller, Gilberto London School of Economics Environmental decision-making often involve multiple impacts and conflicting objectives. Decisions in this context are always complex due to several features: (i) their public nature requires transparency and justification, (ii) the presence of multiple stakeholders with different values, (iii) the need for legitimacy entails participation in the decision-making process, (iv) the existence of a spatial distribution of the different impacts increases the technical complexity of the evaluation, and, finally, (v) there are uncertainties concerning the spatial distribution of impacts and policy makers’ preferences, which must be considered in the analysis. These challenges may explain a growing interest on the integration of Multiple Criteria Decision Aiding (MCDA) with Geographic Information Systems (GIS), as it provides a more efficient way of supporting such decision problems. It allows both the generation and the comparison of spatial alternatives, taking into account policy makers’ conflicting objectives and their preferences as well as spatial information. The scientific literature contains many contributions on MCDA-GIS developments and applications in a variety of domains and several MCDA methods are currently used in order to aggregate the different dimensions of the analysis. In this paper we are considering best practices for spatial multicriteria evaluation and exploring to which extent different popular methods employed in practice comply with these best-practice standards. The review is organized along the key decision analytic phases (model structuring, preference elicitation, aggregation of partial performances, sensitivity analysis and overall communication).We attempt to provide a critical and balanced analysis of the use of the different MCDA methods in the environmental decision-making context, and suggest some opportunities for further developments in this field. Keywords: Decision Analysis, Environmental Decision Making, Multi-Attribute Utility Theory, Multi-Criteria Decision Making, Preference Aggregation, Problem Structuring, Spatial Multi-Criteria Evaluation, Value Focus Thinking. 131 TH3E (invited) Bi-Criteria Optimisation using Surrogate Modelling for Dynamic Process Design Fraga, Eric (*) University College London Beck, Joakim University College London The goal in multi-criteria optimisation is to identify a set of Pareto optimal points that are well spread out in the objective space, with as few evaluations of the computer model as possible. In the context of computationally expensive models, replacing the model with a surrogate in the optimisation cycle may be necessary to solve these problems. Starting with a low-fidelity surrogate model, the surrogate is updated after each iteration of the optimisation. This sequential update procedure generally requires fewer evaluations of the full model than starting with a high fidelity surrogate. Efficient global optimisation (EGO) with Gaussian Process (GP) surrogate models has been widely used for single-objective optimisation. EGO uses the “expected improvement” criterion to select a promising point to update the surrogate model. Recently, Emmerich et al. (2011) used the concept of hypervolume to formulate a generalisation of the expected improvement criterion to multi-criteria optimisation. For single objective optimisation, another criterion is the “expected information gain”, a measure of the conditional mutual information between the points that have been selected and the ones that have not. A good approximation of the expected information gain is available in GP surrogate models, where the prior knowledge is in the points selected so far. Curiosity driven optimisation involves a trade-off between the expected improvement and the expected information gain. This paper investigates different ways of defining information gain in the context of bi-criteria optimisation and explores their impact on the spacing, richness, spread, and coverage of the Pareto front. The use of the expected information gain to update the GP surrogate model for bi-criteria optimisation is demonstrated for the design of pressure swing adsorption processes, maximising the conflicting criteria of purity and recovery of the product. Keywords: Multi-Objective Optimization, Statistics, Surrogate Modeling. 132 TH4G (contributed) Bi-objective Traffic Assignment with Multiple User Classes: A Time Surplus Approach Perederieieva, Olga (*) The University of Auckland Ehrgott, Matthias Lancaster University Raith, Andrea The University of Auckland Wang, Judith Y. T. University of Leeds Traffic congestion is an issue in most cities worldwide. One way to model and analyse the effect of congestion on route choice behaviour is traffic assignment (TA). Conventional TA models the behaviour of travellers by assuming that all drivers are selfish and tend to choose the fastest routes from their origin to their destination. As a result an equilibrium state is achieved, when no one has an incentive to switch to another route. Therefore, it is assumed that all drivers make their travel decisions regarding travel time only. However, this is not true in general. According to empirical studies other important factors are travel time reliability and monetary cost. In the literature on the TA problem where two or more route choice objectives are explicitly distinguished the majority of the models form a weighted sum of the objectives. This approach allows to find only a subset of all equilibrium solutions. We propose to use a conceptually different approach - the bi-objective user equilibrium. It considers two objectives separately (travel time and toll) and allows multiple solutions. In order to model user preferences, we apply the time-surplus maximisation model (TSMaxBUE) which can identify one of the solutions. Time surplus is defined as the maximum time a user is willing to spend minus the actual time spent. The maximum time a user is willing to spend is modelled as an indifference curve - a non-linear function that depends on the path toll. All drivers are divided into classes each with its own indifference curve that models preferences in this group. We show that this model admits a mathematical programming formulation and we adapt some conventional TA algorithms to solve TSMaxBUE with multiple user classes. The TSMaxBUE model allows to obtain various traffic patterns by changing the indifference curves. We observe that this framework is general enough to cover any situation with a flow dependent and a flow independent component of path cost. Keywords: Bi-Objective Traffic Assignment, Time Surplus, User Equilibrium. 133 MO3E (contributed) Biases and Path Dependency in the Even Swaps Method H¨am¨al¨ainen, Raimo (*) Aalto University Lahtinen, Tuomas Aalto University Even Swaps is a multicriteria decision support method introduced by Hammond et al. (1999) in their popular book Smart Choices. In the method, a decision maker carries out a sequence of trade-offs to identify the most preferred alternative out of a set of alternatives. Each sequence that the DM follows creates a path for the decision process. In principle, each path should lead to the same choice. We conducted an experimental study with the SMART-SWAPS software by H¨am¨al¨ainen et al. (2003) to see if this is the case in practice too. The results show that path dependency does exist. We explain how the phenomenon relates to the scale compatibility bias, which causes the dimension that is adjusted in a trade-off question, to receive extra weight (Delqui´e 1993). Besides the scale compatibility bias, we suggest that also the loss aversion may play a role. We argue that the sequential steps in the Even Swaps process can accumulate the biases on some paths, which leads to path dependency. We suggest that it is advisable to do the process repeatedly and check if different paths lead to the same choices. Here computer support provided by the SMART-SWAPS software can help in practice as it allows easy backtracking of the process. References: Delqui´e, P. 1993. Inconsistent Trade-offs between Attributes: New Evidence in Preference Assessment Biases. Management Science, 39(11), 1382-1395. Hammond, J. S., Keeney, R. L., Raiffa, H. 1999. Smart Choices. A Practical Guide to Making Better Decisions. Harward Business School Press, Boston, MA. H¨am¨al¨ainen, R. P., Mustajoki, J., Alanaatu, P., Karttunen, V., Arstila, A. 2003. Smart-Swaps – Smart Choices with Even Swaps. Computer software, Systems Analysis Laboratory, Aalto University, http://www.smart-swaps.hut.fi. Keywords: Bias, Decision Analysis, Decision Support, Even Swaps, Loss Aversion, Scale Compatibility, Smart-Swaps. 134 WE2F (invited) Bidirectional versus Unidirectional Heuristic Search for Multiojective Optimization in State Space Graphs Galand, Lucie (*) LAMSADE - University Paris-Dauphine Ismaili, Anisse Perny, Patrice LIP6 - UPMC Spanjaard, Olivier University Pierre et Marie Curie Multiobjective heuristic search in state space graphs generally aims at determining the exact set of Pareto-optimal cost vectors associated to solution paths or an approximation of this Pareto set. When a preference model is available under the form of scalarizing function f, the search effort merely focuses on the direct determination of an f-optimal tradeoff within the Pareto-set. The solution algorithms proposed in this area rely on unidirectional approaches that develop a front of Pareto-optimal labels attached to subpaths from the initial node s to the goal nodes. When there is a single goal node t, explicitly known, and when search operators are reversible, a bidirectional search approach is possible, interleaving and interacting searches, forward from s and backward from t. Although the respective advantages of unidirectional approaches versus bidirectionnal approaches have been widely discussed in the singleobjective case, they have not been investigated in the setting of multiobjective search. The aim of this paper is to propose a first implementation of multiobjective bidirectional heuristic search and to empirically assess his value compared to unidirectional search. Our comparison concerns both the determination of the entire Pareto set and the determination of OWA-optimal tradeoffs within the Pareto set. We provide results of numerical tests performed on random instances of path-planning problems, both in terms of number of nodes expanded and in terms of computation times. Keywords: Bidirectional Heuristic Search, Multi-Objective Optimization, Multi-Objective Shortest Path Problem, OWA Operator. 135 MO5H (poster) Cautious Solutions in Multi-Scenario Bargaining. An Application to Union-Firm Negotiation. Rubiales , Victoriana (*) University of Seville M´armol, Amparo Mar´ıa University of Seville Monroy, Luisa University of Seville In this paper we address a general two-person bargaining problem under uncertainty. To this end, several states of nature or future scenarios are considered. We propose a solution concept based on the distance to a utopian minimum outcome vector, which guarantees conservative levels of achievement for the agents. This cautious solution is obtained as the solution of a minimax programming problem. An axiomatic characterization of the cautious solution is provided, when some requirement on the bargaining set are assumed. An extension of the classic model of firm-union negotiation, which includes situations where uncertainty about the consequences of the agreements has to be taken into account, is analyzed in this framework. Keywords: Bargaining, Multi-Scenario, Solution Concepts. 136 MO4C (invited) Characterization and Properties of Approximate Proper Solutions in Vector Optimization Problems Huerga , Lidia (*) Jim´enez, Bienvenido UNED Novo, Vicente Guti´errez, C´esar Universidad de Valladolid We present a new concept of approximate proper solution for a constrained vector optimization problem defined on a locally convex Hausdorff topological linear space. We characterize this set of approximate proper solutions through approximate solutions of associated scalar optimization problems by assuming generalized convexity assumptions and we study its properties. Keywords: Linear Scalarization, Nearly Subconvexlikeness, Proper Epsilon-Efficiency, Vector Optimization. 137 FR3G (invited) Classification of Disasters Effects and Humanitarian Decision-Makers’ Criteria Rodriguez, Tinguaro (*) Montero, Javier Vitoriano, Bego˜ na Universidad Complutense de Madrid A fully precise numerical initial assessment of disasters’ effects on people as well as on their social and natural environment is unrealistic in the time-pressured, highly uncertain decision context taking place just after a disaster strike. Instead of a numerical evaluation, here it is considered that it is rather more plausible and realistic to classify the severity of the consequences of a disaster in terms of the relevant scenarios for NGO’s decision makers. However, some features of that context, as the natural ordering of the consequences’ gravity or the need of avoiding underestimation risk, entail the necessity of considering and assuming an structure over the set of classes, somehow modeling those features inside of the classification model. Keywords: Classification, Fuzzy Sets, Humanitarian Logistics. 138 TU5C (contributed) Climate Change - a Challenge for MCDA? Belton, Valerie (*) University of Strathclyde The UNEP website describes climate change as one of the major challenges of our time, giving rise to impacts that are global in scope and unprecedented in scale. It advocates anticipatory action today as future adaptation will become more difficult and costly. The UNEP sponsored MCA4climate initiative provides a framework to assist governments in preparing climate change mitigation and adaptation strategies. At the heart of the framework is a generic MCDA model which provides the basis for the evaluation and prioritisation of potential policy actions. The paper will describe the process of model development, which drew on in depth expertise in 12 mitigation and adaptation themes and incorporated 3 illustrative case studies, and discuss the challenges to MCDA posed by this ambitious initiative. A collaborative project, led by the INCAE Business School in Costa Rica, is currently underway to use the MCA4climate framework to explore the adaptation of the agricultural sector in Peru; progress, challenges faced and hopefully successes achieved with be reported at the conference. 139 TU4B (contributed) Combined Dominance-based Rough Set Approach and Fuzzy DEMATEL for Analyzing the Adoption Intention of Tourist Guide Mobile Applications Chin, Yang-Chieh (*) Asia University The multiple capacity resources and advances of smartphones enhance the demand of mobile applications. Therefore, mobile applications are one of the most rapidly growing segments of the software market. The goal of this project is to gain insight into the factors that affect user intention to adopt tourist guide mobile applications. The present project uses an extended technology acceptance model and focuses on combining innovativeness, intrinsic motivation (i.e., perceived enjoyment) and extrinsic motivation (i.e., perceived ease of use and perceived usefulness) to explain usage intentions for tourist guide mobile applications. Firstly, a dominance-based rough set approach (DRSA), a rule-based decision-making technique, is used to determine the adoption intentions associated with decision rules in toruist guide mobile applications. The second phase of the project uses Fuzzy Decision Making Trial and Evaluation Laboratory (Fuzzy DEMATEL) technique for analyzing the relationship among factors of adoption intention in tourist guide mobile applications. These findings help to address real-world dependence and feedback problems and can help organizational decision-makers in planning, evaluating and executing the use of tourist guide mobile applications. Practical and research implications are also offered. Keywords: Analytic Network Process, DEMATEL, E-Commerce, Fuzzy Integral, Rough Sets, Service Quality. 140 TU5B (contributed) Combining MCDA and Risk Analysis: Dealing with Scaling Issues in the Multiplicative AHP Barfod, Michael Bruhn (*) Technical University of Denmark Salling, Kim Bang Technical University of Denmark Van Den Honert, Rob Macquarie University This paper proposes a new decision support system (DSS) for applying risk analysis and stochastic simulation to the multiplicative AHP in order to deal with issues concerning the progression factors. The multiplicative AHP makes use of direct rating on a logarithmic scale, and for √ this purpose the progression factor 2 is used for calculating scores of alternatives and 2 for calculation of criteria weights when transforming the verbal judgments stemming from pair wise comparisons. However, depending on the decision context, the decision-makers aversion towards risk, etc., it is most likely that the decision-makers have varying progression factors in their perception of the scales. This potential variability in progression factors implies a form of uncertainty in the DSS which can be modelled by stochastic simulation. With reference to previous work, which proposes to conduct sensitivity analysis using a short, natural, and long scale, the proposed paper suggests a new approach in order to capture the influence from the progression factors in terms of including probability distributions. Herein, the uncertainty both with regard to the scale and the inherent randomness from the factor will be captured by probabilistic input and output distributions. Provided that each alternative and criteria under consideration are independent it can be assumed that the embedded uncertainty from the progression factors remains the same. The result will then be an interval estimate (instead of a single point estimate) for each alternative’s final scores. This might lead to overlapping intervals of scores which may be interpreted as possible rank reversals. Thus it will be possible to calculate the probability of overlapping due to the uncertainty in the progression factor for any given set of pairwise comparisons. Simulations on real-life data sets are used to form the conclusions and perspectives for the proposed DSS. Keywords: Analytic Hierarchy Process, Decision Analysis, Decision Support Systems, Risk Analysis. 141 FR2B (contributed) Comparative Analysis and Evaluation of Multi-Criteria Decision Making Methods Kou, Gang (*) UESTC Ergu, Daji Southwest University for Nationalities Saaty, Thomas L. Katz Graduate School of Business Real life decision problems usually involve multiple and conflicting criteria. Many multicriteria decision making (MCDM) methods have been proposed to deal with the measurement of tangible/intangible conflicting criteria and with the measurement of the alternatives of a decision with respect to these criteria in order to choose the best one between them. However, the existence of many MCDM methods becomes a decision problem, and decision makers may be uncertain about which one to use. Thus it is also important to compare and evaluate the existing MCDM methods, and explore if there are logical or mathematical, social or practical reasons why one method is better than another. This paper first evaluates a simple decision choosing the best car among three cars, using several (more than ten) MCDM methods including the AHP, Utility, VIKOR, PROMETHEE II, DEMATEL, ELECTRE, ARAS, COPRASS, MACBETHE, UTA etc. Based on this validation, we are further concerned with the comparison and evaluation of these different methods on a wide set of criteria such as 1) Good and complete understanding of the importance of a decision; 2) The Creativity, Generality, Breadth and Depth of Decision Structure ; 3) Faithfulness of judgments (Individual and Group); 4) Scale of measurement; 5) The way of obtaining weights; 6) Measurement of intangibles; 7) Applicability to conflict resolution; 8) Social political group decision making (GDM); 9) Identifying incompatible decision maker in the context of GDM; 10) How to combine judgment; 11) How to deal with uncertainty; 12) Applicability to feedback; 13) Validity of the outcome (the ability to predict). The goal of this paper is to analyze the similarities and differences of each MCDM method and summarize the common characteristics of each to create a general framework for multi-criteria decision making methods. The analyzed results can help decision makers to choose the appropriate multi-criteria decision making method to make a valid decision. Keywords: Comparative Analysis, Conflicting Criteria, Multi-Criteria Decision Making. 142 TH2D (invited) Comparing Fuzzy Goal Programming Approaches for Collaborative Supply Chain Master Planning Diaz-Madro˜ nero, Manuel (*) CIGIP-UPV Mula, Josefa Universitat Polit`ecnica de Val`encia Peidro, David Universitat Polit`ecnica de Val`encia The objective of this paper is improving the proposal by Selim et al. (2008) Collaborative production-distribution planning in supply chain: A fuzzy goal programming approach. Transportation Research Part E: Logistics and Transportation Review, 44(3), 396-419. In the context of a supply chain collaborative planning problem, the authors differentiate that an approach based uniquely in the weighted sum of the decision maker satisfaction levels can be corresponded to a centralized collaborative perspective while an approach based on the min operator can be corresponded to a decentralized one. In this paper, we compare the original proposal by Selim et al. (2008) to other representative fuzzy goal programming approaches for decentralized and/or centralized master planning problem in a real-world ceramic tile supply chain. This work has been funded by the Universitat Polit`ecnica de Val`encia projects: ‘Material Requirement Planning Fourth Generation (MRPIV)’ (Ref. PAID-05-12). Keywords: Collaborative Planning, Fuzzy Sets, Goal Programming, Meta-Goal, Production, Supply Chain Management. 143 TU5E (contributed) Comparing the Scalarization Methods for Bi-Objective Assembly Line Scheduling Problem Akyol, Emine (*) Anadolu University Sara¸c, Tugba Eskisehir Osmangazi University In this study, bi-objective assembly line scheduling problem in a plastic part production process is considered. The objective functions are minimized the sum of tardiness and the sum of earliness. The objective functions are scalarized by using well-known scalarization methods in the literature and obtained results are compared. Keywords: Assembly Line, Scalarization, Scheduling. 144 TH3G (invited) Comparison between Continuous and Discrete Multicriteria Decision Methods applied to Lands Use, Conservation, Management and Planning on Hydro-Basin in Cordoba Province (Argentina) Grau, Juan (*) Andina, Diego Ant´on, Jos´e Tarquis, Ana Universidad Polit´ecnica de Madrid Cisneros, Jos´e Manuel Universidad Nacional de R´ıo Cuarto The authors have attempted to study the similarities and differences between the application of continuous and discrete Multi-criteria methods to the maintenance and basin hydrological management and Planning, taking into account the land use of the influence area and its characteristics. Firstly, the authors have obtained data on soils and on hydrology of the study area, and evaluated the systems of soil uses and actions that could be recommended considering the relevant aspects of the study area. Several characteristics such as: crops, erosion, employment, economic sustainability, etc. have been considered. Later on, Decision Support Systems (DSS) with Mathematic tools for planning of defences and uses of soils in these areas have been applied, and these were conducted here to use multi-criteria decision models (MCDM), first discrete, for the more global views about soil conservation and hydraulic management actions and to chose among global types of use of soils, and second continuous, in order to evaluate and optimize combined actions, compare the results obtained with discrete methods and to see if they could provide another significant aspects of the problem, such as, control of erosion and hydraulic management. For that, they used discrete methods for the different sub-basins, ELECTRE, weighted and original PROMETHEE and AHP methods with a system of criteria grouped as environmental, economic and social, to obtain a better view on the general aspects for the middle or long term planning for these areas. In relation to continuous methods, they used Weighted and Lexicographic Goal Programming and Compromise Programming. This contribution shows the results obtained, the discussion about them and the use of one or another, and finally the recommendation to apply MCDM to hydrobasin Management and Planning. Keywords: Applications of MCDM, Case Study, Environment, Multi-Criteria Decision Making. 145 TU4C (contributed) Comparison of Two Methods for Developing a Multicriteria Evaluation System to Assess Animal Welfare Mart´ın Fern´andez, Paula (*) CAU Kiel Buxad´e, Carlos Polytechnic University Madrid Krieter, Joachim CAU Kiel The aim of this paper is to develop a multicriteria evaluation system to assess animal R distinguishes four welfare criteria (Good feeding, Good houswelfare. Welfare Quality ing, Good health and Appropriate behaviour). A farm receives one score for each welfare criteria. Nine farms were used as an example illustrated on fattening pigs. Multiattribute utility theory was used to produce an overall value of welfare starting from the data regarding the four main criteria. The utility functions and the aggregation function were constructed in two separated steps. First, utility functions for each criterion were determined in two different ways, by using the standard sequences method (SS) and by using the MACBETH software. In the second step, criteria were aggregated. As aggregation functions, the weighted sum (WS) and the Choquet Integral (CI) were used. The WS was implemented with different weights to analyse the sensitivity of the method. A progressive interactive approach was used in the CI determination to define the relative importance of the different criteria and the allowance or not of compensation between criteria. Results showed that when using the WS the ranking derived from the utility functions calculated by the MACBETH approach was more consistent with the preferences of the decision maker (DM) than the ranking derived from the SS method, but, for both methods the weights have a high influence on the results, having several ranking reversal when the weights are slightly modified. From the CI results we could notice that there were interactions between the criteria, therefore, the use of the WS in this context is not adequate due to the fact that it does not take into account these interactions. The utilities derived from MACBETH suited better with the progressive interactive approach proposed to calculate the CI, allowing to model better the preferences of the DM regarding the different importance of the criteria and the interaction between them. Keywords: Animal Welfare, Choquet Integral, Criterion Importance, Multi-Attribute Utility Theory, Sensitivity Analysis, Weighted Sum. 146 TU5F (contributed) Computing Nadir Point for Multiobjective Discrete Optimization Problems Kirlik, Gokhan (*) Koc University Sayin, Serpil Koc University Optimization problems with more than one objective function is called a multiobjective optimization problem (MOP). In this paper, we investigate the problem of finding the nadir point for multiobjective discrete optimization problems (MODO). MODO is a special case of MOP where all variables are discrete. The nadir point is characterized by the componentwise worst values of objective functions among nondominated points of a MOP. Obtaining the nadir point is generally a hard problem. Along with the relatively easy to obtain ideal point, the nadir point is an important element of MOP, because these points define lower and upper bounds of the efficient set. In fact, there are some methods that require the nadir point as input. Hence, determination of the nadir point is an important concept in MOP. We present a new algorithm to compute nadir values for MODO with p objective functions. The proposed algorithm is based on an exhaustive search of the (p-2)-dimensional space for each component of the nadir point. We partition this space into (p-2)-dimensional rectangles to search the space entirely. During the search, epsilonconstraint scalarization is used to obtain efficient solutions. Since the optimal solution of the epsilon-constrained method is weakly efficient, we use second stage to eliminate those solutions. The method guarantees to find the nadir point in a finite number of iterations. We compare our algorithm with an earlier one from the literature. Both algorithms are implemented in C++. Subproblems are solved by using IBM CPLEX 12.4. These algorithms are tested on the randomly generated multiobjective knapsack problem, assignment problem, and integer linear programming problem instances. Our algorithm is able to obtain the nadir point for relatively large problem instances with up to fiveobjectives. Keywords: Multi-Objective Optimization, Nadir Point. 147 MO3G (invited) Concept and Empirical Verification of the Benefit of Value-Focused Thinking Kunz, Reinhard (*) University of Bayreuth Siebert, Johannes University of Bayreuth Value-focused Thinking is attracting more and more attention in literature and in practice. Its great popularity indicates that a decision maker can benefit from using Value-focused Thinking. An extensive literature review shows that previous studies concentrate on the results and not on the process of Value-focused Thinking. Le´on (1999) found out that using Value-focused Thinking leads to a more extensive hierarchical structure of objectives in comparison to Alternative-focused Thinking. Thereupon the structure generated by Value-focused Thinking is evaluated at least equal or superior by experts. Selart and Johanson (2011) explore that Value-focused Thinking seems to have a positive impact on the quality of ideas in terms of creativity and innovativeness. In this paper we present a concept to measure the benefits using different perspectives and considering different aspects of Value-focused Thinking. Several constructs and valid management scales are identified and empirically tested in a pre-study measuring the benefits of Value-focused Thinking. Keywords: Decision Analysis, Measurement, Problem Structuring, Value Focus Thinking. 148 MO2D (contributed) Constructing Ecological Value Functions for River Rehabilitation: Expert Elicitation with Different Aggregation Schemes Lienert, Judit (*) Eawag Langhans, Simone Eawag Reichert, Peter Eawag River rehabilitation aims to reduce human impacts on e.g. biodiversity and ecosystem services, but is expensive and entails difficult trade-offs. Although suitable, MCDA has not received much attention as a tool to support selecting appropriate rehabilitation measures. Many applications were simplified (few objectives, linear value functions, additive aggregation model). Additional challenges include the best choice of attributes for valuation. We propose a comprehensive ecological value function to assess the biological and hydromorphological state of smaller rivers in Central Europe. In 6 one-to-one interviews with river ecologists, we developed a large objectives hierarchy, elicited value functions for over 60(!) attributes, scaling constants (weights), and each ecologist’s risk attitude. We discussed implications of an additive model and suggest alternative schemes (minimal, multiplicative, cobb-douglas, mixed aggregation). We set up hypothetical scenarios (“decision alternatives”) for attribute states and calculated the overall value of the river’s ecological state, based on each expert’s preferences. We then used different aggregation schemes, and assumptions about scaling (e.g. equal weights), or value functions (linear) to test the sensitivity of the result. We discuss that although hierarchical multiplicative aggregation proved satisfactory in the interviews, mixed aggregation may better fulfill desired properties of an ecological value function. The large number of attributes can be turned into an advantage with hierarchical aggregation from lower to higher objectivelevels. It increases flexibility (application to various river types and spatial scales) and allows selecting preferred attributes. If more attributes are used, the statistical significance of branches increases without changing weights. Our findings may be relevant for many practical MCDA-applications with extensive objectives hierarchies and non-linear, and/or non-compensatory attributes. Keywords: Aggregation Schemes, Decision Maker’s Preferences, Elicitation, Environmental Decision Making, Multi-Attribute Utility Theory, Water Resources Management. 149 MO3G (invited) Creating a Balanced Scorecard by using Value-Focused Thinking Muetterlein, Joschka (*) University of Bayreuth Kunz, Reinhard University of Bayreuth Siebert, Johannes University of Bayreuth Since its introduction by Kaplan and Norton in 1992 the Balanced Scorecard has started a remarkable triumphal march. These days the majorities of the companies in the Dow Jones use the Balanced Scorecard to measure and control their performance. The widespread use manifests itself also through thousands of articles and books dealing with extensions and further developments. It is therefore surprising that the procedure how to create a Balanced Scorecard is given only little attention. In literature there are often only practical guides which describe in detail how in workshops of the top management and of the lower management have to be organized to define the vision, mission and strategy of the company. In contrast their systematic implementation in different perspectives and consistent translation in performance indicators under consideration of cause-and-effectnetworks is only insufficiently addressed. We propose an innovative procedure to identify and structure the relevant objectives using Value-focused Thinking (Keeney 1992). Fundamental objectives can be used to create the perspectives in a Balanced Scorecard. Instead of cause-and-effect-networks means-end-networks represent the relations between the performance indicators within or in between the perspectives. This procedure ensures a systematic and consistent implementation of the perspectives and performance indicators. We illustrate our procedure with a case study. Keywords: Case Study, Decision Analysis, Measurement, Problem Structuring, Value Focus Thinking. 150 MO1A (plenary) Creating, Structuring, and Using Multiple Objectives Keeney, Ralph (*) Duke University For an important decision, a set of objectives provides a logical basis for all of the activities necessary to solve that decision. The first part of this presentation discusses the difficulties of identifying new objectives, and then presents procedures that significantly help in creating an appropriate set of objectives for any decision. Then concepts and procedures to structure these objectives are outlined. The second part of the presentation describes several uses of the objectives. These include framing a stated, or even just perceived, decision problem for a subsequent analysis, creating alternatives to solve a decision, recognizing proactive decision opportunities that would contribute to the achievement of sets of objectives, prioritizing objectives by constructing a multiple-objective objective function, and evaluating alternatives proposed for a decision to identify the best one or better ones and eliminate poor choices. All of the procedures to create the structure multiple objectives and to use them are illustrated with personal, business, and public applications. Keywords: Applications of MCDM, Creating Alternatives, Identifying Objectives, Organizing Objectives, Prioritizing Objectives, Structuring Decisions. 151 FR2F (invited) Cross Entropy for Combinatorial Optimization Problems with Linear Relaxations Molina Luque, Juli´an (*) University of M´alaga Caballero Fern´andez, Rafael University of Malaga Hern´andez-D´ıaz, Alfredo G. Pablo de Olavide University Laguna, Manuel University of Colorado While the cross entropy methodology has been applied to a fair number of combinatorial optimization problems with a single objective, its adaptation to multiobjective optimization has been sporadic. We develop a multiobjective optimization cross entropy (MOCE) procedure for combinatorial optimization problems for which there is a linear relaxation (obtained by ignoring the integrality restrictions) that can be solved in polynomial time. The presence of a relaxation that can be solved with modest computational time is an important characteristic of the problems under consideration because our procedure is designed to exploit relaxed solutions. This is done with a strategy that divides the objective function space into areas and a mechanism that seeds these areas with relaxed solutions. Our main interest is to tackle problems whose solutions are represented by binary variables and whose relaxation is a linear program. Our tests with multiobjective knapsack problems and multiobjective assignment problems show the merit of the proposed procedure. Keywords: Cross Entropy, Evolutionary Multi-Objective Optimization, Multi-Objective Assignment, Multi-Objective Combinatorial Optimization, Multi-Objective Knapsack. 152 TU2F (contributed) Dealing with Migratory Flows under Uncertainty: a Multi-objective Optimization Approach Campoy-Mu˜ noz, Pilar (*) Universidad Loyola Andaluc´ıa Garc´ıa-Alonso, Carlos R. Universidad Loyola Andalucia Salazar-Ordo˜ nez, Melania Universidad Loyola Andaluc´ıa Nowadays, managing migratory flows has become one of the most relevant challenges for politicians in the recipient countries. However, the estimation of those inflows has turned into a complex problem given both the plethora of variables driving them and the multiple interactions among drivers.This paper presents a tool to help governments forecast immigration flows by combining simulation and fuzzy logic techniques. This approach allows dealing with the uncertainty arisen from the scarce data available, but also results into a multi-objective optimization problem regarding to the fitting of the model parameters. To tackle it, an evolutionary algorithm is designed for this task. The proposal is used to evaluate the Ecuadorian inflows to Spain during the last decade. The results are consistent with those previously obtained by expert-knowledge elicitation while providing some useful insights for national migratory policy design. Keywords: Fuzzy Logic, Genetic Algorithms, Immigration, Monte Carlo Simulation, Multi-Objective Optimization. 153 MO5H (poster) Decision Making using Immune Algorithm for Increasing of Level Crossing Capacity Gorobetz, Mikhail (*) Riga Technical University The purpose of this research is multi-objective optimization of level-crossing traffic capacity by decision making using immune algorithm. There are conflicts between different transport types in cities, where road and rail transport is mixed. Road transport is forced to wait at level crossing, while train is passing. Current level-crossing control system does not take in account criteria of train location and speed. And the existing schedule shows that opposite directions trains do not pass the level-crossing simultaneously. By these reasons, a level-crossing is closed longer than might be. Therefore some solutions to optimize the schedule and decision making for safe and efficient closure of level-crossing are necessary. Hypothesis, that the existing schedule adjustment and embedded devices may reduce the time of level-crossing in a closed position, is defined. In this research the mathematical model of level-crossing railway and road transport flows is created taking in account specifics of railway control and signalization systems and target function is developed for multi-objective optimization by criteria of road transport idle time, simultaneity, safety and shift from the original schedule. Dual population immune algorithm is used for schedule optimization. Also genetic algorithm has been used with the same target function and model to compare the results. The computer model as well as the 24h schedule of real existing problematic level-crossing with heavy transport traffic has been optimized and evaluated. Statistics is collected from various computer experiments with genetic and immune algorithms. Developed transport system control algorithm allows increasing the level crossing capacity up to 49% for computer model and up to 25% for real levelcrossing. The research leading to these results has received funding from the ARTEMIS Joint Undertaking under grant agreement n◦ 295373 and from Latvian Academy of Science and Ministry of Education and Science. Keywords: Immune Algorithm, Multi-Objective Optimization, Traffic. 154 TU3G (contributed) Decision Support for Multi-Attribute Multi-Item Reverse Auctions Karakaya, G¨ ul¸sah (*) Middle East Technical University Koksalan, Murat Middle East Technical University In this study, we address multi-item auction problems in a multi-attribute, multi round reverse auction setting. In single-attribute auctions there is one attribute, typically the price. In multi-attribute auctions there are additional attributes and the comparison of bids is not straightforward. Multi-item auctions also bring additional complexity over single-item auctions. In the single-item auction, the winning bidder supplies the item with the committed attribute values. On the other hand, in multi-item case it is not so trivial to determine the winning bidders. In such auctions, generally bidders offer a combination of items – a bundle – they wish to supply. They specify the attribute values of the bundle and make their bids. Multi-item auctions are known as combinatorial auctions. The auction process continues several rounds. In each round, we provide the buyer with a set of efficient bid combinations. The buyer then chooses the provisional winners whose bids collectively provide all the required items. We estimate preference information from the buyer’s past choices and provide this information to the bidders. The bidders update/improve their bids in order to become/stay competitive. The aim of the developed interactive approach is to have the more competitive bidders eventually end up winning the auction. We tried the algorithm for underlying linear preference functions. We also adapted the approach as a heuristic for the case the buyer has a nonlinear preference function. There, we locally approximate the buyer’s preference function with a linear function. The approach worked well for the problems we solved. Keywords: Auctions/Bidding, Decision Support, Interactive Methods, Multi-Criteria Decision Making. 155 TH4B (contributed) Decomposing Value Creation when Assessing Investments: a Multi-Criteria Approach Guerrero-Baena, Mar´ıa Dolores (*) University of C´ordoba G´omez-Lim´on, Jos´e A. University of Cordoba Although traditional financial techniques based on the NPV and the IRR are widely used by corporate officers in order to support decision-making for capital budgeting, many researchers have criticized them for failing to recognize the real value generated by new investments, since they ignore qualitative impacts that also add value to the firm. In this paper, we aim at partially solving this shortcoming by developing a novel approach to appraising investments considering the overall value created by the new assets, which can be decomposed into financial and nonfinancial value, the latter stemming from intellectual capital. To do so, a methodology based on the Analytic Hierarchy Process (AHP) is proposed, which allow assessing more accurately the value creation of the different alternative investments and supporting a more adequate decision-making. The method proposed is empirically illustrated by a case study taken from the agrifood sector in order to check its operational feasibility, where an investment in a food quality control technology is considered. Keywords: Analytic Hierarchy Process, Capital Budgeting, Intellectual Capital. 156 TU2B (contributed) Deriving Priority Weights from Pairwise Comparison Matrices under Different Rationality Scenarios Romero, Carlos (*) Technical University of Madrid Gonz´alez-Pach´on, Jacinto Tecnical University of Madrid The final purpose of the pairwise comparison method is to deduce a set of numerical values representing the relative importance of different objects for a decision maker (DM). In this paper, this task is tackled by considering different scenarios of rationality. In the literature this problem has been addressed by two directions: when we state “a priori” hypothesis about the rationality of the DM, or when the rationality study is made “a posteriori”; i.e., when a complete pairwise comparison matrix has already been obtained. In the first direction, we consider three different scenarios according to the level of rationality underlying the information provided by the DM: rational decision-maker, partially rational decision-maker and irrational decision-maker. For the second direction, we consider again three different approaches: normative, descriptive and a compromise between normative and descriptive approaches. Keywords: Distance-Based Methods, Goal Programming, Pairwise Comparison, Preference Modeling. 157 TH4C (contributed) Design Safer and Greener Road Projects by using a Multi-Objective Optimization Approach Sarrazin, Renaud (*) Universite libre de Bruxelles De Smet, Yves Universit´e libre de Bruxelles For many years, considering sustainable development and improving road safety have been two majors concerns in mobility and transport policies in Europe. Since 2001, the European Commission (EC) had published several reports about European transport in which actions have been taken to improve the global level of safety on the European road network by 2020. In 2010, the Greening Transport Package had been published about strategies to apply in order to strive for a transport system more respectful of the environment. But to date, the assessment of the road safety of an infrastructure is essentially based on the evaluation of statistical data that offers the administration a support in the identification of the roads with high accidents concentration – or black spots. These methods are based on a reactive approach. However, to meet the objectives of the EC of improving road safety and considering sustainable character of the road transport infrastructure, it has become essential to develop new preventive and innovative tools. Consequently, we had initiated this research project with the aim of developing a multicriteria analysis methodology to carry out an integrated and preventive assessment of the road safety performances and the sustainable aspects of road projects at the design stage. At first, we have developed an innovative concept of sustainable safety that we transformed into a complete set of criteria. And then, we have studied the interest of using a multi-objective optimization (MOO) methodology to conduct a pertinent analysis of our design problem. This work illustrates the efficiency of MOO in design space exploration and the development of interesting solutions. Moreover, the use of MOO allows us to deal with several criteria and then to design alternatives more adapted to the stakes and the characteristics of a road project. In the end, these results would help the design engineers in the development and the selection of safer and greener solutions. Keywords: Multi-Objective Optimization, Road Safety, Sustainability Assessment. 158 TU4E (contributed) Determinants of Experienced Utility: Laws and Implications Sarin, Rakesh (*) UCLA Baucells, Manel Rand Corporation Satisfaction in experiencing the future depends on decisions made today. We consider six well-known psychological laws governing satisfaction. The laws capture habit formation, social comparison, and satiation. We show it is possible to formalize these laws by means of a utility model, and to derive implications from the laws: wanting vs. liking, crescendo, recharge periods, variety seeking, and craving. The discussion combines mathematical propositions, experimental findings in psychology, and time-honored wisdom. We discuss how the sixth law, presentism, may lead to incorrect predictions of experienced utility and suboptimal life-balance choices. Keywords: Decision Analysis, Utility Theory. 159 FR3G (invited) Development of an Innovative Environmental Education Program for Urban Solid Waste Management through Multi-Criteria Decision Methods (MCDM) Grau, Juan (*) Andina, Diego Ant´on, Jos´e Tarquis, Ana Universidad Polit´ecnica de Madrid Martin, Rodrigo UNED Soletti, Alfredo Universidad Cat´olica de C´ordoba This contribution shows the design of an innovative program of Environmental Education for the integrated management of urban solid waste (USW) with the aid of discrete multi-criteria decision methods. In the development and elaboration of strategies, the authors have taken into account all the levels to which it is addressed, the socio-economic conditions of the area. They have also considered the main objective of creating a culture oriented to progressive change of habits and practices in the efficient management of USW from the place of origin up to its final disposal through its storage and transport. Keywords: Applications of MCDM, Multi-Criteria Decision Making, Outranking Methods. 160 MO3C (contributed) Development of Interactive Support System Using Spreadsheet based on Multiattribute Utility Analysis Tomohiro, Hayashida (*) Hiroshima University Ichiro, Nishizaki Hiroshima University Mitsuhiro, Imai Hiroshima University Multiattribute utility analysis (MAUA) (Keeney and Raiffa, 1976) is a multi-criteria decision method, such that preference structure of a decision maker is quantitatively-modeled by multiattribute utility function. The function consists of multiple weighted single attribute utility functions, the weight of each single attribute is indicated by a scale constant. A decision support system based on MAUA “MIDASS” is developed by Seo et al. in 2007. This system makes analysts avoid complicated calculations under MAUA. MIDASS consists of several applications having particular functions for identifying the single attribute utility functions and the scale constants, calculation of expected utility value of each alternative, and database management. It can operate on windows OS. MAUA is an effective approach for a wide range of decision problems, however, MIDASS is supposed that an expert user at whole procedure of MAUA. Due to complexity of operations of MIDASS, the software is difficult for a user who is not an expert. In this study, we develop an interactive decision support system which runs on spreadsheet with basic functions and easy to operate for multiattribute utility analysis. Keywords: Decision Support, Multi-Attribute Utility Theory, Multiplicative Form. 161 FR3G (invited) Development of Mathematical Multicriteria Methods to Aid to Elaborate Plans of Rural Sustainable Development in Areas With Aborigines Communities Ant´on, Jos´e (*) Andina, Diego Grau, Juan Tarquis, Ana Universidad Polit´ecnica de Madrid Rubio, Jos´e Luis Universidad de Valencia The purpose of this paper is to present the development of mathematical models based on discrete multi-criteria methods to be used in the development of rural management plans in municipalities with social integration problems, especially when aborigine communities are settled in their territory. In these municipalities there also exist problems of natural hazards and environmental erosion, desertification and adaptation to possible climate change. All this is accompanied by serious economic shortcomings that move to lack of planning scenarios for sustainable economic development of the territory. So that the design of a method to aid decision making in the development plans to be carried out by the municipalities and regions is of high importance in most of the Latin American countries. Keywords: Environment, Multi-Criteria Decision Making, Rural Planning. 162 FR3G (invited) Discussion of the Results Obtained in Several Real Applications using PROMETHE I and II versus Weighted PROMETHE and other Outranking Methods Grau, Juan (*) Andina, Diego Ant´on, Jos´e Tarquis, Ana Universidad Polit´ecnica de Madrid Colombo, Federico De los R´ıos, Lisandro UCASAL The authors have carried out a deep analysis of the results obtained in these four real applications: 1) Optimization of an erosion control integral plan in Chaco area in Salta Province (Argentina), 2) Territorial planning in La Colacha river basin (C´ordoba Province, Argentina), 3) Election of water resources management entity, 4) Election of waste management system in an Andean village using several discrete multi-criteria decision methods. PROMETHEE I, II and a PROMETHEE modified by the authors named weighted PROMETHEE have been used. Also, others outranking methods as ELECTRE I and AHP have been tested. This contribution shows the reasons why the weighted PROMETHEE should be used in place to I and II for some kind of applications. Keywords: Applications of MCDM, Outranking Methods, PROMETHEE. 163 TH2E (invited) Distributed Computation of Pareto Points in Multiobjective Programming Wiecek, Margaret (*) Clemson University Dandurand, Brian Clemson University Complex systems engineering design has motivated the development of optimization methods dealing with multiple design disciplines, with each discipline generating multiple design criteria. The complexity is reflected in the system composition of subsystems and components. Multiple disciplines originate from various science and engineering areas, such as structures or dynamics, which interact with each other within the design process, while multiple criteria are required to describe the system performance. Interaction among the design processes of the multiple disciplines is necessary due to the natural overlap and influence that the underlying scientific principles have on one another. To address the needs of engineering design, the design process is modeled as the Multidisciplinary Multiobjective Optimization Problem (MMOP), and interdisciplinary communication protocols and algorithms for distributed computation of Pareto points of the MMOP are developed. The MMOP is presented as distinct and nonintegrable multiobjective subproblems corresponding to the interacting disciplines. The protocols govern the interactions between the subproblem solvers during the design process, enable specialization, guarantee stability of the system-wide design process, and coordinate the computation of system-wide Pareto designs. The algorithms make use of two Gauss-Seidel-type decompositions known as Block Coordinate Descent (BCD) and the Alternating Direction Method of Multipliers (ADMM), and a Lagrange multiplier method known as the method of multipliers (MM). BCD and ADMM are applied to MMOPs whose decompositions originate in the decision and objective space, respectively. The convergence properties of the algorithms, which had been established in the single objective case, are examined in the multiobjective setting. Mathematical and engineering examples are included. Keywords: Block Coordinate Descent, Decomposition, Lagrangian Duality, Multi-Objective Optimization. 164 MO5H (poster) Distribution of Federal Funds to Mexican Subnational Governments: Identifying Tools to Enhance MCDA Process in the Public Security Sector Vergara-Maldonado, E. Ivonne (*) S´anchez-Guerrero, Gabriel National Autonomous University of Mexico This research concerns the identification of tools and methods that can be integrated into a framework to aid the process of decision making related to the distribution of federal funds to the Mexican states, in the public security sector. Since violence and crime have risen over the past decade, the Federal Government increased considerably the annual federal budget in order to allocate public funds across the national territory to prevent, combat and penalize the crime. One of these resources is the Fund of Contributions for Public Security, which contributed with about 7 billion of Mexican pesos in 2012. This budget is distributed to 32 states based on a model with five main criteria, a considerable number of sub-criteria, and weighted sums. Since its establishment in 1997, this model has been frequently questioned and modified in their criteria and weights. The last change was in 2010, and it was derived from a participatory process with stakeholders from the National System of Public Security; where the great number of variables in the social phenomenon of crime, legal framework, political constraints and a wide variety of interests due the complexity of the system, gave as a result an unfeasible model, which needed an urgent correction in order to meet the restrictions from the main decision-maker, the National Council of Public Security. This research analyzes this situation in an agreement/certainty framework (Patton, 2011) and identifies the appropriate multi-criteria approach to manage this type of decision making processes (Munda, 2004). Finally, using the typology of systems methodologies related to problem context (Flood & Jackson, 1995; Jackson, 2003), this work suggests a few methods and tools that can be integrated into a MCDA framework, in order to enhance the viability and acceptance of the design and results of the model to distribute this public fund. Keywords: Federal Funds Distribution, Multi-Criteria Decision Aid, Multi-Criteria Decision Analysis, Multi-Methodology. 165 MO5H (poster) DR-MOEA/D: A new Double Reference Point Approach of MOEA/D Based in Aspiration and Reservation Levels Saborido-Infantes, Rub´en (*) University of M´alaga Alba, Enrique University of Malaga Luque, Mariano University of Malaga Evolutionary multiobjective optimization (EMO) methodologies have attained wide popularity for solving practical optimization problems, due to their ability to find a representative set of Pareto-optimal solutions. The multiple criteria decision making (MCDM) literature also offers many possibilities to carry out this task involving user preferences, which can be supplied in different forms. We propose a variant of the popular MultiObjective Evolutionary Algorithm based on Decomposition (MOEA/D), called DR-MOEA/D, for finding a preferred well distributed reduced set of solutions to approximate a region of interest, instead of the complete Pareto-optimal frontier, by incorporating preference information of the decision maker based on the information included in two reference points (aspiration and reservation). Using these reference levels, we generate a well distributed set of reference points in the region of interest, covering the objective space from the aspiration level. Considering these reference points, an achievement scalarizing function (ASF) is used to project solutions on the non-dominated set. DR-MOEA/D approximates the region of interest delimited by two reference points (aspiration and reservation), and gets a set of Pareto-optimal solutions, if the aspiration point is unachievable and the reservation point is achievable. In other cases, when there no Pareto-optimal solutions exist in the region of interest, the algorithm projects this region onto the Pareto-optimal set. The method is validated with several standard problems of the literature (families CTP, ZDT, DTLZ and WFG). Keywords: Achievement Scalarizing Functions, Evolutionary Optimization, Multi-Objective Optimization, Pareto Optimal Solutions, Preference-Based Evolutionary Algorithms. 166 WE1G (contributed) Dynamic Multi-Objective De Novo Linear Programming Fiala, Petr (*) University of Economics Prague Mathematical programming under multiple objectives has emerged as a powerful tool to assist in the process of searching for decisions which best satisfy a multitude of conflicting objectives. There are a number of distinct methodologies for multi-objective decisionmaking problems that exist. In Multi-objective Linear Programming (MOLP) problems it is usually impossible to optimize all objectives in a given system. Trade-off means that one cannot increase the level of satisfaction for an objective without decreasing this for another objective. Trade-offs are properties of inadequately designed system a thus can be eliminated through designing better one. Multi-objective De Novo linear programming (MODNLP) is problem for designing optimal system by reshaping the feasible set by given prices of resources and a given budget. Many of today’s systems operate in a dynamic environment. The paper presents approaches for solving dynamic MODNLP problems. Dynamic MODNLP as an extension of MODNLP can deal with time dependent changes in systems. The models try to reflect changes in real or simulated time and take into account that the model components are constantly evolving. Keywords: De Novo Programming, Dynamic Approach, Multi-Objective Optimization. 167 MO2F (invited) Dynamic Resampling for Guided Evolutionary Multi-Objective Optimization of Stochastic Systems Siegmund, Florian (*) University of Sk¨ovde Deb, Kalyanmoy Michigan State University Karlsson, Alexander University of Sk¨ovde Ng, Amos University of Sk¨ovde In Multi-objective Optimization many solutions have to be evaluated in order to provide the decision maker with a diverse Pareto-front. In Simulation-based Optimization the number of optimization function evaluations is very limited. If preference information is available however, the available function evaluations can be used more effectively by guiding the optimization towards interesting, preferred regions. One such algorithm for guided search is the R-NSGA-II algorithm. It takes reference points provided by the decision maker and guides the optimization towards areas of the Pareto-front close to the reference points. In Simulation-based Optimization the modeled systems are often stochastic and a reliable quality assessment of system configurations by resampling requires many simulation runs. Therefore optimization practitioners make use of dynamic resampling algorithms that distribute the available function evaluations intelligently on the solutions to be evaluated. Criteria for sampling allocation can be a.o. objective value variability, closeness to the Pareto-front indicated by elapsed time, or the dominance relations between different solutions based on distances between objective vectors and their variability. In our work we combine R-NSGA-II with several resampling algorithms based on the above mentioned criteria. Due to the preference information R-NSGA-II has fitness information based on distance to reference points at its disposal. We propose a resampling strategy that allocates more samples to solutions close to a reference point. Previously, we proposed extensions of R-NSGA-II that adapt algorithm parameters like population size, population diversity, or the strength of the Pareto-dominance relation continuously to optimization problem characteristics. We show how resampling algorithms can be integrated with those extensions. The applicability of the proposed algorithms is shown in a case study of an industrial production line for car manufacturing. Keywords: Evolutionary Multi-Objective Optimization, guided search, Reference Point, Resampling, Simulation-Based Optimization, Stochastic Systems. 168 TU3E (invited) E-NAUTILUS: A Surrogate based Interactive Multi-objective Optimisation Method without Trading Off Sindhya, Karthik (*) Miettinen, Kaisa University of Jyv¨askyl¨a Luque, Mariano Ruiz, Ana Bel´en Ruiz, Francisco University of Malaga Interactive methods are commonly used in multiobjective optimization, where the decision maker (DM) participates in the solution process by iteratively providing preference information to find her/his preferred Pareto optimal solution(s). Recently, a new interactive multiobjective optimization method called NAUTILUS has been proposed. At each iteration of NAUTILUS, a new solution is shown to the DM, which dominates the previous one. This avoids the necessity for the DM to trade-off between objectives and only the last solution is Pareto optimal. We utilize the philosophy of the NAUTILUS method and propose here a new method called E-NAUTILUS. In the E-NAUTILUS method we generate a surrogate for the original multiobjective optimization problem using a representation of Pareto optimal solutions, before starting the interactive solution process with the DM. This enables us to obtain new solutions whenever desired without repeatedly solving the original multiobjective optimization problem. For this reason, E-NAUTILUS is especially useful for computationally expensive problems. Any suitable (evolutionary) multiobjective optimization algorithm is used to generate a set of nondominated solutions and these solutions are used to construct a surrogate of the original problem. As in the NAUTILUS method, the solution process is started from the nadir point and at every iteration a clustering scheme is used to select and show several solutions (which dominate the solution of the previous iteration) to the DM, who expresses her/his preferences of how objective function values should be improved by choosing one. Finally, when the DM has found her/his preferred solution (which is a solution to the surrogate problem), it is projected onto the actual Pareto optimal front. We illustrate the working principle of the E-NAUTILUS method using a problem related to electrical industry. Keywords: Evolutionary Multi-Objective Optimization, Multi-Objective Optimization, Nadir Point, Preference Information, Reference Point. 169 TH3F (contributed) Efficient Optimisation of a Vehicle Suspension by Hybrid Nature Inspiring Method Syeda, Darakhshan Jabeen (*) Indian Statistical Institute In this paper, we propose a new hybrid method based on region reduction division criteria (RRDC) and advanced real coded genetic algorithm (ARCGA) for solving the suspension design problem. The problem is concerned with determining the passive suspension parameters of the vehicle for comfortable ride. The two dimensional model of the vehicle with non-linear suspension springs and linear dampers has been considered. Moreover, in the dynamical model two passengers in the seated position, one at the front and the other at the rear position of the vehicle have been considered. To enhance comfort of the passengers technological constraints as per ISO standards has been introduced to the optimization problem. The suspension parameters have been found by minimizing the bouncing transmissibility of the sprung mass subjected to the constraints in time domain. In order to solve the problem, the constrained optimization problem is first converted into an unconstrained one by penalty function technique. Then to the transformed problem the proposed hybrid method is applied. To verify the validity of the obtained parameters, model simulation has been done over different roads and the results obtained have been compared with the results obtained with the existing suspension parameters. Keywords: Artificial Intelligence, Control, Genetic Algorithms, Penalty Methods. 170 MO4C (invited) Efficient Solutions for Non-Regular Multiobjective Problems Hern´andez Jim´enez, Mar´ıa Beatriz (*) Universidad Pablo de Olavide Osuna-G´omez, Rafaela Universidad de Sevilla Ruiz-Garz´on, Gabriel Universidad de C´adiz Necessary optimality conditions for efficient solutions are presented for multiobjective problems with inequality-type constraints. These conditions are applied when the constraints do not necessarily satisfy any regularity assumptions, and they are based on the concept of 2-regularity introduced by Izmailov. In general, the necessary optimality conditions are not sufficient and the efficient solution set is not the same as the KarushKuhn-Tucker points set. So it is necessary to introduce generalized convexity notions. In the multiobjective nonregular case we give the notion of 2-KKT-pseudoinvex-II problems. This new concept of generalized convexity is both necessary and sufficient to guarantee the characterization of all efficient solutions based on the optimality conditions. Keywords: Non-Regular Problems, Optimality Conditions. 171 WE1B (contributed) R The AHP Project on the Cloud Elephas: P´erez-Rodr´ıguez, Fernando (*) Rojo-Alboreca, Alberto University of Santiago de Compostela ´ Camino-Saco, Angeles G´omez-Garc´ıa, Esteban VSonCloud Mosquera, Alejandro In management that we understand how Sustainable should to consider criteria like Economic, Social, Environmental and Technical properly. In an analytical assessment process which involves multiple criteria and alternatives the decision process is complicated. Besides the criteria may have two types: objective and subjective. The objective criteria are easily quantifiable, and regardless of the complexity of their evaluation, you can always do a rigorous statistical analysis or methodologies like optimization by Gauss-Jordan, while the subjective criteria are associated with uncertainty due to its subjective nature or opinion. In the analysis of subjective criteria must apply methodologies that associate to each weight or size. These methodologies are collectively referred to as multicriteria analysis, and although each has its own synthesis calculation, have in common that the decision was made based on a comparison emitted by a / some decision maker / s and under a predetermined scale. One of the methodologies used and studied is the Analytic Hierarchy Process (AHP) proposed by Saaty in the early eighties. This methodology is based on the allocation of weights to the criteria and / or alternatives through pairwise comparisons, i.e. a decision maker evaluates their preference between two competing criteria under a given scale. With the growing popularity of cloud computing, SaaS applications (Software as a Service) is becoming the software delivery model of choice for organizations around the world. With SaaS, software is available in the cloud, allowing users to access the Internet using a thin client such as a web browser. Among the advantages of SaaS can highlight its rapid deployment, lower maintenance, reduced costs and ease of adding new features as well as the possibility of changes in real time. In this paper we present Elephas R an implementation of cloud-based AHP free access for making decisions. , Keywords: Analytic Hierarchy Process, Computing Science, Criteria, Decision Support Systems, Pairwise Comparison. 172 MO4F (contributed) Elicitation Procedures for the Comparison of Decisional Maps Brison, Valerie (*) Universit´e de Mons Pirlot, Marc Universit´e de Mons In the field of land management, it is not uncommon to have a geographic map, representing a region under study, and divided into geographic units. Each unit is assessed on an ordinal scale describing its degree of suitability for some usage, for instance housing, or its state of degradation with respect to sustainable development criteria. We call such a map a decisional map. After a while, and for example, after the application of some policies aiming to improve the situation, the state of the units evolves. For some units, the state has improved, and for some others, it has deteriorated. What we want to know is whether the global state of the map has improved or not. The purpose of this work is to provide models to help a decision maker to compare the state of a region at different stages of its evolution. We formulate two mathematical models and we provide an interactive elicitation procedure to determine all the parameters of the models developed. For this purpose, we transpose a widely used method in the theory of decision under risk, namely the comparison of lotteries. We interpret lotteries as maps and we formulate questions to the decision maker in terms of comparisons of well-chosen maps. Keywords: Elicitation, Map Comparison, Preferences. 173 TU5G (invited) Engineering Multiobjective Shortest Path Heuristics Zaroliagis, Christos (*) CTI & University of Patras Multiobjective shortest path is a core problem within multiobjective optimization, appearing in applications such as QoS routing in communication networks, traffic equilibria, transport optimization as well as route and itinerary planning. It is an extension of the single criterion shortest path problem using a cost vector (instead of a scalar) per network edge, representing multiple (often conflicting) objectives or criteria. Even though numerous efficient algorithms exist for the single criterion shortest path problem, the multiobjective counterpart of the problem is much harder (NP-hard), since the optimal solution (captured by the well-known concept of the Pareto set) is typically exponential in size. Two main approaches are followed in order to reduce the computation effort for solving the multiobjective shortest path problem. The first one uses approximation and computes optimal solutions up to a certain factor. Approximation techniques do not necessarily yield exact solutions, but are adequately fast to be used in practice. The second approach uses heuristic improvements to speed-up existing algorithms. Such techniques yield exact solutions, but it is considerably more difficult to achieve good performance. In this work we focus on the latter approach, motivated by the great demand in practical applications to achieve efficient and exact multiobjective shortest paths. Building upon a new graph structure specifically suited for large-scale networks, we present new implementations of heuristic algorithms (including NAMOA*, the currently best such heuristic) for the solution of the multiobjective shortest path problem. We enhance the heuristics with further optimizations and experimentally evaluate the performance of our enhanced implementation on real world road networks achieving 10 times better performance with respect to the best previous study. Keywords: A* Search, Heuristic Search, Multi-Objective Shortest Path Problem. 174 TU5G (invited) Engineering Parallel Bi-Criteria Shortest Path Search Erb, Stephan (*) Karlsruhe Institute of Technology Sanders and Mandow (IPDPS 2013) present a parallel label-setting algorithm for the multi-criteria shortest path problem. The algorithm is a generalization of Dijkstra’s single-criteria shortest path algorithm for the multi-criteria case. Using a Pareto queue, a multi-dimensional generalization of a priority queue, the algorithm is able to parallelize all additional work when going from one to multiple criteria. For the bi-criteria case it can even be shown that the algorithm performs less work than previous sequential algorithms. We discuss in how far these theoretical results translate into efficient implementations for modern shared memory multiprocessors. In particular, we focus on a cache-efficient implementation of the bi-criteria case and present an extensive evaluation of the algorithm and its underlying data structures. Keywords: Multi-Objective Optimization, Parallel Computing, Performance Measurement, Shortest Path. 175 MO4G (contributed) Enhancing Portfolio Risk Analysis - Balancing the Past, the Future, the Objective, the Subjective Myers, Chris (*) The University of Manchester Risk analysis is an important element of the decision making process of insurance companies. Market participants such as capital providers, rating agencies and regulators recognize this importance, and as a result often want transparency from insurers as to their risk profiles. This transparency may include reporting on extreme risk measures such as value-at-risk or tail value-at-risk with respect to investment portfolios. Most methods rely heavily, if not exclusively, on historical risk and return data to calculate and forecast loss impacts given stress environments. However, historical risk and return characteristics are far from a perfect or reliable indicator of the future, particularly under extreme environments. Moreover, blindly following any quantitative modeling framework without subjective overlays can make for misleading conclusions regarding the risk profiles that these frameworks were meant to articulate. This discussion will show examples of the potential limitations of risk analysis that is purely quantitative in nature, and will introduce an approach that blends historical and forecasted risk and return assumptions for improved risk analysis and decision making. For example, one can use historical information to understand certain useful characteristics that define the shape of the distribution of returns for invested assets. This understanding could be enhanced by overlaying current and prospective views of those returns, perhaps supported by forward looking economic indicators, which would define the body of that same distribution. Thus one has an analytical approach that balances objective and subjective decision analysis. The focus of the discussion will be towards investment portfolios typical of non-life insurance companies, but may very well be appropriate for other investment strategies. Keywords: Decision Support, Finance, Investment Analysis, Risk Analysis. 176 TU4B (contributed) Enhancing Smart Mobile Devices by the Fuzzy DEMATEL based Network Process Tzeng, Gwo-Hshiung (*) Kainan University Huang, Chi-Yo Kao, Yu-Sheng National Taiwan Normal University The smart mobile device market surged dramatically as the embedded system and wireless communication technology advances during the past decade. As a result, people depend heavily on smart mobile devices for retrieving information ubiquitously. Therefore, efficient computation capability is essential for smart mobile devices. In order to explore the factors influencing customers’ acceptances of smart mobile communication technologies and enhance the less efficient features accordingly, a fuzzy multiple criteria decision making (FMCDM) framework will be defined. The fuzzy Decision Making Trial and Evaluation Laboratory (DEMATEL) based Network Process (DNP) will be used to construct the casual relationship and derive weights being associated with the factors. The factors with higher weights are the more influential ones. Meanwhile, the performance scores and the satisfactory level corresponding to the factors as well as the scores of competing smart mobile devices targeting at the same market segmentation will be measured based on performance measurement APPs as well as experts’ opinions. The factors with lower performance scores or satisfactory level will be enhanced by appropriate innovation strategies being selected based on the Grey Relational Analysis. An empirical study on an Android based smart mobile device being provided by a world leading Taiwanese IT vendor will be used to verify the feasibility of the proposed framework. The analytic results being derived by using the proposed Fuzzy DNP based framework methods can be the basis for IT managers and marketers’ strategy definitions. Keywords: DEMATEL-Based ANP, Fuzzy MCDM, Innovation Strategy, New Product Development, Smart Devices. 177 TU2E (invited) Entropy Measures to Control Robustness in Ordinal Regression Models Greco, Salvatore (*) University of Catania Siskos, Yannis University of Piraeus Slowinski, Roman Poznan University of Technology Within Multiple Criteria Decision Aiding (MCDA), ordinal regression methods build a preference model in the form of a value function or an outranking relation on the basis of some preference information provided by the Decision Maker (DM). Since very often there exists a plurality of instances that represent equally well the preference information given by the DM, Robust Ordinal Regression (ROR) has been proposed with the aim of considering the whole set of compatible instances of a chosen preference model representing the preference information given by the DM. ROR provides a recommendation of the MCDA procedure in terms of possible preference, in case the preference holds for at least one compatible instance, and necessary preference, in case the preference holds for all compatible instances. Using ROR there is a risk that the necessary and possible preferences are too vague because the set of compatible instances is “too large”. In order to control this aspect of the ROR methodology, we propose to measure the set of compatible instances of the considered preference as the hypervolume of the set of corresponding preferential parameters. Moreover, we propose to build a measure of the uncertainty due to ambiguous representation of DM’s preferences based on a credibility of preference obtained from application of Stochastic Multiobjective Acceptability Analysis (SMAA) methodology to ROR. The certainty about the truth or falsity of the preferences can be modeled through this credibility of preferences using the concept of Shannon entropy. Entropy measures as well as the hypervolume of the sets of preference parameters compatible with DM’s preference information can be used to assist the process of elicitation of preferences by the DM, in order to evaluate the tradeoff between the decreasing confidence of the preference information given by the DM and the reduction of the hypervolume of compatible preference parameter set and the improvement of the entropy measures. Keywords: Entropy, Multi-Criteria Decision Aid, Ordinal Regression, Robust Ordinal Regression, Stochastic Multi-Objective Acceptability. 178 MO2E (contributed) Error Correction Model in Classification by Using Multiple-Criteria and Multiple-Constraint Levels Linear Programming Wang, Bo (*) Xianhua, Wei University of Chinese Academy Sciences Shi, Yong UESTC Nowadays, optimization based classification methods are widely researched, support vector machine (SVM) for example. In addition to this, based on goal programming (GP) and multiple-criteria linear programming (MCLP), various classification methods according to linear system are posed and developed. And then, considering multiple-constraint levels, some researchers proposed algorithms by using multiple-criteria and multiple-constraint levels linear programming (MC2LP). Until now, a system of linear models, MCLP and MC2LP, has been applied to classification problems. However, these previous methods cannot help increasing the accuracy of classification directly. That is to say, there is no method focuses on minimizing the misclassified cases, which is a key issue in evaluating of a certain classification algorithm. Furthermore, in Statistics, misclassified cases can be seen as two types of error. At the same time, MCLP model finds a solution to project all the points into a one dimension space so that the value of cutoff can discriminate which group every point belongs. According to this fact, the correction of two types of error can be obtained through changing the cutoff of the classification hyperplane. Therefore, we expect to make use of MC2LP to get a changeable cutoff, which has a special connection with the right hand side of the constraints. In this situation, a brand new MC2LP based model, which can be applied to deal with two types of error respectively, can be developed. In this paper, we pose a new MC2LP model that can correct two types of error. Especially, it can be developed into a pair of models, which can deal with two types of error respectively. Besides, we will give the matrix representation of the new models and discuss some properties of them. At last, some experiment results based on the new models will be displayed to show the new model’s capability in correcting two types of error. Keywords: Matrix Representation, Multi-Constraint Levels, Multi-Objective Linear Programming, Multi-Objective Optimization, Two Types Of Error. 179 WE2G (invited) Estimating the Pareto Front of a Hard Bi-Criterion Competitive Location Problem Lancinskas, Algirdas (*) Vilnius University Fern´andez Hern´andez, Pascual Universidad de Murcia Pelegr´ın, Blas Universidad de Murcia Zilinskas, Julius Vilnius University We dealt with a location problem for an expanding franchise type firm in competition with other firms in a geographical area. The aim of the firm is maximizing the market share captured by the new facilities and minimizing the the lost of market share of the old facilities caused by the entering of the new facilities in the market. The market share of each facility is estimated assuming that customers are served by the most attractive facility. We introduce a new tie breaking rule to serve the customers for which there are more than one facility with maximum attraction which lead to solve a bi-objective hard nonlinear optimization problem. An heuristic algorithm is proposed which let obtain a good approximation of the Pareto front when the firm have to select a fixed number of new facilities in a finite set of candidates. Keywords: Evolutionary Multi-Objective Optimization, Location. 180 TH2C (contributed) Evaluating Clinical Performance of Pediatric Emergency Physicians with a Help of Data Envelopment Analysis Model Michalowski, Wojtek (*) Javier, Fiallos University of Ottawa Patrick, Jonatha Farion, Ken Children’s Hospital of Eastern Ontario The clinical performance of physicians working at a Pediatric Emergency Department (PED) is a key factor in determining the ability of PED to provide efficient and quality patient care. The assessment of PED physicians is a multiple criteria task involving such measures as the patient health outcomes, the timeliness of care, the throughput of patients, and the efficient use of resources, however most of the literature concerned with physician performance evaluation uses a single criterion such as medical competency or quality of care. There were few attempts at multi-criteria evaluation of physicians’ performance and all of them applied some form of Data Envelopment Analysis (DEA). None of these evaluations consider criteria related to the quality of health outcomes and the timeliness of care. The DEA model proposed here focuses on the clinical performance of PED physicians. This performance is assessed using a criteria related to diagnostic effectiveness, timeliness of care, throughput of patients by individual PED physicians and quality of health outcomes measured for these patients. These four criteria capture the multi-faceted nature of the PED physicians’ clinical work and, to the best of our knowledge, it is the first time when such a comprehensive evaluation of clinical performance has been proposed. A DEA model known as the Slacks Based Measure model was selected as the evaluation method. This model allows measuring the excess of inputs and shortfall of outputs that characterize an inefficient PED physician’s performance. It calculates a score that captures these inefficiencies, identifying benchmark physicians. While use of the model assumes that the physicians have control over both their inputs and outputs, model’s solution supports the development of improvement plans for each inefficient physician. The model was developed using data reflecting operations of PED in one of Ottawa (Canada) teaching hospitals. Keywords: Clinical Performance, Data Envelopment Analysis, Healthcare, Multi-Criteria Decision Making. 181 MO3E (contributed) Evidential Reasoning Rule for Criteria Combination in MCDM Xu, Ling (*) The University of Manchester Yang, Jian-Bo The University of Manchester A unique Evidential Reasoning (ER) rule is established for combining multiple criteria with various weights and reliabilities in decision and judgement making. The novel concept of Weighted Belief Distribution (WBD) is proposed and extended to WBD with Reliability (WBDR) to complement Belief Distribution (BD) as introduced in DempsterShafer (D-S) theory of evidence. The implementation of the orthogonal sum operation on WBDs (WBDRs) leads to the establishment of the new ER rule, which constitutes a generic conjunctive probabilistic reasoning process, or a Bayesian inference process when used to combine probability information. It is shown that the ER combination algorithm developed for multiple criteria decision analysis under hybrid uncertainty is a special case of the ER rule when the reliability of information is reflected by its weight and the weights of all criteria are normalised. It is proven that Dempster’s rule is also a special case of the ER rule when each piece of information is fully reliable, with the latter completing and enhancing the former by identifying how to combine multiple pieces of fully reliable information that are highly or completely conflicting through a so-called reliability perturbation analysis. The main properties of the ER rule are explored to facilitate its application. Some typical existing rules are discussed and compared with the ER rule. Numerical and simulation studies are conducted to show the validity and features of the ER rule. Keywords: Bayesian Inference, Belief Distribution, Dempster-Shafer Theory, Evidential Reasoning, Multi-Criteria Decision Analysis. 182 FR2G (contributed) Experimental Evaluation of Polynomial and Copula Functions for Qualitative Option Ranking Bohanec, Marko (*) Joˇzef Stefan Institute Mileva-Boshkoska, Biljana Joˇzef Stefan Institute We address the problematics of full option ranking in MCDA. We start from qualitative models created according to DEX methodology, which ensures option classification into preferentially ordered classes. To obtain option ranking within each of the classes, the QQ method is used as an extension to DEX. QQ constructs a family of linear regression functions, each one corresponding to a different class. The main drawback of QQ is that it frequently leads to partial ranking or fails to satisfy the monotonicity of options within classes. Hence we modify QQ to investigate the usage of polynomial and copula-based regression functions for option ranking. In order to find the best polynomial function for a given set of options we use the CIPER algorithm. CIPER uses a specific heuristics to define and search the space of possible polynomial functions for the given set of options. As a results it finds one or several polynomial functions that satisfy the heuristics and that provide best fit for the data. Copulas are functions that capture the non-linear dependences among random variables. To use copulas for option ranking, each attribute is considered as a random variable. The variables are nested into hierarchical bi-variate copula structures to determine the non-linear dependences among all attributes at hand. The hierarchical copula structures are used to obtain regression functions for option ranking. In this paper we empirically evaluate the proposed methods on three groups of experiments with two, three and four input attributes and a class attribute. We show that polynomial functions manage to provide best results of fulfilling the monotonicity in the first two experiments, however they lead to partial ranking. Finally, copulas manage to provide best results in terms of satisfying both the monotonicity and full option ranking in all three groups of experiments. Keywords: Copula-Based Model, Linear Model, Multi-Criteria Ranking, Polynomial Model, Qualitative Multi-Criteria Model. 183 WE1F (invited) Explicit Equations for Non-Dominated Frontier in a Portfolio Selection Problem based on Soft Computing Salda˜ na, Alberto Carlos (*) Cadenas, Jos´e Manuel Garrido, Maria del Carmen Martinez, Raquel University of Murcia The more natural approach to the portfolio selection problem has the double objective of maximizing the return on an investment while minimizing risk. The set of risk-return pairs of all the efficient portfolios is called non-dominated frontier, and so the decision-support techniques designed to assist an investor in these problems consist of computing and analyzing this frontier. We address two kinds the constraints which make the corresponding model more involved from a computational viewpoint: semicontinuous variable constraints and cardinality constraints. The non-dominated frontiers of such problems become more irregular and new specific computation techniques are required. We describe a multiobjective GA that allows us to obtain the non-dominated frontier of the proposed portfolio problem through explicit expressions of the arcs of parabola. ENCODING: Individuals are binary strings. In addition, each individual has associated seven values: (a,b,c) which describe the parabola and intervals [rmin,rmax], [Rmin,Rmax] in which the parabola is defined. FITNESS: We define the fitness as the relative Level of Non-Dominance (LND): an individual ’y’ has a LND value obtained from an interval [A,B] into [Rymin,Rymax] where its parabola (a,b,c) dominates to any other parabola corresponding to other individual ’y1’. SELECTION AND GENETIC OPERATORS: a) We incorporate elitism. The best individuals are those with a fitness value of LND=1. b) The selection is by means of binary tournament and due to the structure of individual, we define a special binary tournament from the fitness function and the distance of niches. c) Due to the individual structure we define an hybrid crossover. d) The used mutation is the classical mutation of a gene. The output is the set of individuals with LND other than 0. These individuals describe the non-dominated frontier through explicit expressions of the arcs of parabola. From these expressions all values of portfolios are obtained. Keywords: Multi-Objective Genetic Algorithms, Non-Dominated Frontier, Portfolio Selection, Soft-Computing. 184 MO3G (invited) Facing the Energy Transition: An Illustration how Value-Focused Thinking helps to Find the Right Answers Krampf, Peter (*) University of Bayreuth Siebert, Johannes University of Bayreuth Germany is facing a huge disruptive change in its energy system. The speed of the transition from the fossil fuel and nuclear energy age to the renewable energies had been tremendously accelerated after Fukushima and the government´s decision to phase out of nuclear energy in the next 10 years. To ensure long-term competitiveness, to define the future strategy and to agree on the right implementation steps, currently traditional German utilities face decisions with huge consequences. In the following 20 years decentralized renewable energy solutions, cost intensive offshore wind parks and the decommissioning of centralized conventional power plants have to be planned and realized. Like in other industries, there is the tendency to deal with decision problems isolated from each other without focusing on the big picture. Thereby it is often only vague defined what and how should be achieved. Furthermore the search of alternatives is based on traditional thinking, which is highly dangerous in a disruptive change. Value-focused Thinking provides methods to identify and structure objectives (what should be achieved). The objectives can be used to create systematically alternatives and decision opportunities (how this could be achieved) which the company was not aware of before. Even known in literature, so far there had been no practical use in the energy sector. In this presentation we demonstrate based on a practical case example how Value-focused Thinking is able to support this challenge one of the top 12 utilities in Europe, located in Germany in linking future decisions explicitly to their objectives. After demonstrating the challenge, we illustrate the identification and structuring of objectives. We show how the objective‘s networks is affected by the energy transition from the fossil fuel and nuclear energy age to the renewable energies. Finally we demonstrate how alternatives could be created in a specific decision problem under consideration of its objectives. Keywords: Case Example, Energy System, Utility, Value Focus Thinking. 185 FR2D (contributed) Factors Affecting Marketing Channel Choice Decisions in Citrus Supply Chain: A Conceptual Framework Siddique, Muhammad Imran (*) Massey University Garnevska, Elena Massey University Marr, Norman Massey University Citrus fruit supply chain is multifaceted and complex due to the presence of a large number of intermediaries. Selection of an appropriate marketing channel by growers involves a complex decision making process. A number of scocio-economic, demographic and psychological factors are involved in this decision making process. This paper reviews the existing factors affecting decision making in agribusiness and models and methods employ to measure and analyse these factors. A conceptual framework based on multi-attribute utility theory is proposed to measure the most important factors affecting the citrus growers and pre-harvest contractors marketing channel choice decisions in Pakistan. A multivariate decision analysis technique, conjoint analysis, is proposed to identify and evaluate the major factors affecting marketing channel choice decisions of citrus growers and pre-harvest contractors. The research would provide valuable information about citrus growers and pre-harvest contractors decision making process and thus would contribute to improving the efficiency and effectiveness of citrus industry. Keywords: Conjoint Analysis, Decision Making, Multi-Attribute Utility Theory, MultiVariate Decision Analysis. 186 TH3C (contributed) Feature Selection for Classification Using Multi-Objective Optimization Cimen, Emre (*) Anadolu University Ozturk, Gurkan Anadolu University Feature extraction is necessary and the key point for classification problems. Especially in real time systems, like object detection systems, features must be extracted from the image rapidly. There is a trade of computation time with accuracy in classification. It is not possible to exctract all features because of time constraint. Also using less features decrease classification accuracy. In this study we use a mathematical programming approach to find classification function with promising features by considering two objectives: number of features and classification accuracy. Keywords: Classification, Feature Selection, Mathematical programming. 187 MO2F (invited) FEMOEA: a Fast and Efficient Multi-Objective Evolutionary Algorithm L´opez-Redondo, Juana (*) University of Granada Fern´andez-Hern´andez, Jos´e University of Murcia Mart´ınez-Ortigosa, Pilar University of Almer´ıa A multi-objective evolutionary algorithm which can be applied to many nonlinear multiobjective optimization problems is proposed. Its aim is to quickly obtain a fixed size set of points approximating the complete Pareto-front. It adapts ideas from different multi-objective optimization evolutionary algorithms, but also incorporates new devices. In particular, the search in the space is carried out on promising areas determined by hyperspheres (subpopulations). Their radii are not fixed; on the contrary, they decrease as the optimization procedure evolves. This mechanism of a decreasing radius helps to maintain a balance between exploration and exploitation of the search space. Additionally, a new local method which accelerates the convergence of the population towards the optimal Pareto-front, has been incorporated. Basically, this method is an extension of the local optimizer SASS to multi-objective problems, which improves a given solution along a search direction (no gradient information is used). Finally, a termination criteria has been also incorporated, which stops the algorithm whenever the changes experimented in the candidate Pareto-front are negligible (less that a given tolerance) during three consecutive iterations. To know how far two approximation sets are from each other, a new Hausdorff-like distance is proposed. In order to analyze the algorithm performance, it has been compared to reference algorithms NSGA-II and SPEA2 on a set of twenty instances, which includes, among others, the ZDT and DTLZ suites of benchmark problems. Several quality indicators have been considered, namely, hypervolume, average distance, additive epsilon indicator, spread and spacing. According to the computational results and statistical analysis performed, the new algorithm, named FEMOEA, outperforms, on average, both NSGA-II and SPEA2 in all the quality indicators. Keywords: Improving Method, Non-Linear Multi-Objective Optimization, Quality Indicators, Stopping Rule. 188 FR3C (contributed) Finding Nadir Points in Multi-objective Integer Programs Lokman, Banu (*) Aalto University Koksalan, Murat Middle East Technical University The nadir value of a criterion is the worst possible value in that criterion over the set of nondominated points. Nadir points have important uses in multiple criteria decision making problems. Knowing the nadir point and the ideal point allows us to properly scale the problem. The nadir point is also a good starting point for algorithms that start from dominated points and try to converge to preferred points. Finding the nadir point is not straightforward when there are three or more criteria. It is particularly hard to find the nadir point for Multi-objective Integer Programs (MOIPs). We develop an exact algorithm that finds the nadir point for any MOIP. In order to find the nadir value of a given criterion in a maximization problem, we define an upper bound for the nadir in that criterion. We also set lower bounds to each criterion by using the properties of the nadir point. We then iteratively generate points in a region defined by these bounds and update the bounds using these points. At each iteration, the algorithm tries to find a nondominated point having a smaller value than the upper bound in the criterion of interest. The algorithm stops when no nondominated point can be found in the searched region. We also develop an approximation algorithm that finds lower and upper bounds for the nadir value in each criterion with a desired level of accuracy. The approximation algorithm stops when the relative gap between the bounds falls below a desired value. We test the performance of the algorithms on several instances of multi-objective assignment, knapsack, and shortest path problems. Our computational experiments on three and four criteria problems show that the algorithms work well. Keywords: Integer Programming, Multi-Objective Optimization, Nadir Point. 189 TH2G (invited) Finding Representative Subsets in Multiobjective Discrete Optimization Paquete, Luis (*) University of Coimbra Ponte, An´ıbal Polytechnic Institute of Set´ ubal Stiglmayr, Michael University of Wuppertal Vaz, Daniel University of Coimbra We consider several formulations of the problem of selecting a representative subset of alternatives from the nondominated set of a biobjective discrete optimization problem, according to some property of interest. These formulations can be seen as a special type of facility-location problems. In particular, we introduce algorithms to find subsets with respect to the epsilon-indicator, coverage, uniformity and combinations thereof. We analyse these algorithms in terms of time and space complexity and discuss their performance in an experimental setting. Extensions to more objectives are also presented. Keywords: Combinatorial Optimization, Multi-Criteria Combinatorial Optimization, Multi-Objective Optimization. 190 TH2G (invited) Finite Representation of Non-Dominated Sets in Multi-Objective Linear Programming Ehrgott, Matthias (*) Lancaster University Shao, Lizhen University of Science and Technology Beijing We address the problem of representing the continuous and non-convex set of nondominated points of a multi-objective linear programme by a finite subset of such points. We prove that a related decision problem is NP-hard. Then we illustrate the drawbacks of the known global shooting, normal boundary intersection, and normal constraint methods concerning the coverage of the non-dominated set and uniformity of the representation by means of examples. We propose an algorithm which combines the positive elements of both the global shooting and normal boundary intersection methods but overcomes their limitations of. We prove that the new algorithm computes a set of evenly distributed non-dominated points for which the the coverage error and the uniformity level can be measured. Finally, we illustrate the usefulness of the new algorithm with an application in the planning of radiation therapy for cancer treatment and present numerical results for randomly generated multi-objective linear programmes. Keywords: Linear Programming, Multi-Objective Optimization, Representative Set. 191 MO3F (contributed) Framing Mechanical Design as Multiple Criteria Decision Making Processes Kaliszewski, Ignacy (*) Systems Research Institute Kiczkowiak, Tomasz Technical University of Koszalin Miroforidis, Janusz Treeffect Co, and Systems Research Institute In the presentation we pursuit our earlier development [1] on coupling local evolutionary computation based search for Pareto front elements with a preference capturing and passing mechanism. This time we approach two mechanical engineering design problems, namely the problems of round tube beam and pneumatic actuator design. We use those two problems, analysed and solved before by classical structural mechanics and optimisation methods, to illustrate the versatility of the development and its natural scalability, mainly trough the use of evolutionary computations, to medium and large scale practical problems. The development offers itself as a reasonable option to all the approaches to MCDM which presume the knowledge of Pareto fronts, or at least the knowledge of fair representations of those sets. We use also those two problems to illustrate an interpretation of the preference capturing and passing mechanism we use as a universal communication interface - a universal language of sort - between humans and computer based multiple criteria optimisation models. We support our views by examples from mechanical engineering literature where needs for such a language/interface to cope with interdisciplinary design challenges have been articulated for long. [1] Kaliszewski I., J. Miroforidis, D. Podkopaev, Interactive Multiple Criteria Decision Making based on preference driven Evolutionary Multiobjective Optimization with controllable accuracy. European Journal of Operational Research, 216, 2012, 188-199. Keywords: Evolutionary Multi-Objective Optimization, Mechanical Engineering, MultiCriteria Decision Making, Multi-Objective Optimization. 192 WE1F (invited) Fuzzy Decentralized Production and Distribution Planning in an Uncertain Environment Diaz-Madro˜ nero, Manuel (*) CIGIP-UPV Mula, Josefa Universitat Polit`ecnica de Val`encia Peidro, David Universitat Polit`ecnica de Val`encia A supply chain (SC) may be considered an integrated process in which a group of several organizations, such as suppliers, producers, distributors and retailers, work together to acquire raw materials with a view to converting them into end products which they distribute to retailers. One way to coordinate operations in this context is by centralized planning, in which a single decision maker has access to all SC members information details (i.e. stocks, production capacities, costs, etc.) in order to create optimal global plans for all partners in the SC. However, although it has been reported that the benefit of information sharing is significant in reducing total SC costs and optimizing capacities, materials and production decisions, SC partners could be reluctant to reveal all of their information trying to locally optimise their own objectives, generating unsuitable or infeasible centralized plans. On the other hand, the competitive and globalized environment and prominent role of customers’ preferences could imply an important degree of uncertainty in production and distribution decisions. It can be distinguished between randomness, or epistemic uncertainty or lack of knowledge of the parameter values. In order to overcome the need for confidential information of all members in SC by a single decision maker and to generate a profitable production and distribution plan under uncertainty conditions we propose a modified approach of the decentralized coordination scheme by Jung and Jeong (2005) with minimal information sharing by incorporating the ranking fuzzy numbers method proposed by Jim´enez (1996) to handle the lack of knowledge in model parameters. This work has been funded by the Universitat Polit`ecnica de Val`encia project: ’Quantitative Models for the Design of Socially Responsible Supply Chains under Uncertainty Conditions. Application of Solution Strategies based on Hybrid Metaheuristics’ (PAID-06-12). Keywords: Collaborative Planning, Distributed Decision Making, Fuzzy Sets, Supply Chain Management, Uncertainty Modelling. 193 TU2D (invited) Fuzzy Linguistic Multi-Criteria Morphological Analysis applied to Scenario Planning Cruz Corona, Carlos (*) University of Granada Lamata, Mar´ıa T. University of Granada Masegosa, Antonio D. University of Granada Villacorta, Pablo University of Granada Nowadays, companies that compete for a place in rapidly evolving markets of modern society know the necessity of good practices that help them to deal with uncertainty, make better decisions in the present -especially regarding management of innovation- and hence anticipate the future. The most employed technique is scenario planning, which consists of a set of tools that emphasize the systematization of creative thinking to uncover possible future scenarios. They consider possibilities that may have escaped to ad-hoc human studies by procedures which capture the most important aspects of both current situation and feasible futures, highlight implications and impact of present decisions, and sketch the steps to go from the present to the desired future. Some of the methodologies proposed to develop this general idea are Cross Impact Analysis, Trend Impact Analysis, Fuzzy Cognitive Maps and Morphological Analysis (MA). We focus on MA, which is aimed at identifying all scenarios that result from the combination of non-quantifiable states that the system variables can take, as well as the feasibility (compatibility) of each combination. An interesting variant of MA is MORPHOL, which also allows evaluating the probability of each scenario by aggregating the probabilities of each of the states adopted by the variables in that scenario. The probabilities mentioned above are numerical values provided by experts based on their own expertise and intuition, so they are subject to certain vagueness not captured by a crisp value. Furthermore, scenarios are judged only in terms of probability but no other criteria are considered. Therefore it is impossible to uncover those scenarios that are the most desirable, catastrophic or feasible, or even the best ones when considering all criteria simultaneously. In this preliminary study, we try to overcome both issues by a novel MCDM procedure of MA that introduces multiple criteria and linguistic fuzzy labels scenario evaluation. Keywords: Applications of MCDM, Computing With Words, Fuzzy Numbers, Linguistic Variables, Morphological Analysis, Scenario Planning, Scenarios, Soft-Computing. 194 TU2D (invited) Fuzzy Meta-Goals in Goal Programming Jim´enez L´opez, Mariano (*) Universidad del Pa´ıs Vasco Arenas-Parra, Mar Universidad de Oviedo Bilbao, Amelia Universidad de Oviedo In goal programming it is not uncommon that the DM does not accept the first solution. In order to address this problem, additional meta-goals, derived from the original set of goals, have been proposed in the literature. The most common meta-goals are related to the percentage sum of unwanted deviations, the maximum percentage deviation and the percentage of achieved goals. In this paper we propose relaxing the last ones by referring to “nearly achieved” goals. This imprecise expression is modeled by a fuzzy set whose membership function represents the degree in which the target is nearly achieved. In the same way the second meta-goals, related to maximum percentage deviations, can be graduated by fuzzy sets. This way we obtain a model simpler than those obtained with crisp meta-goals, whose solution, in addition, could be better balanced. Some illustrative examples are solved. Keywords: Fuzzy Sets, Goal Programming, Meta-Goal. 195 WE1F (invited) Fuzzy Multi-Criteria Decision Models for the Portfolio Selection Problem Vercher, Enriqueta (*) Universidad de Valencia Berm´ udez, Jos´e D. Universidad de Valencia The portfolio selection problem deals with the decision about which assets a portfolio should comprise considering a suitable tradeoff between return and risk. Regarding the quantification of the uncertainty of future returns on a given portfolio we use LR-type fuzzy variables whose expected moments measure the risk and profitability of the investment. We formulate a fuzzy mean-absolute deviation model where investor preferences, portfolio size and diversification conditions are introduced by means of constraints. We solve the corresponding multi-objective optimization problem within the multi-criteria framework by using an evolutionary procedure which provides non-dominated portfolios in the approximated Pareto frontier. To select the best portfolio among the Pareto solutions we apply a ranking strategy based on the concept of fuzzy Value-at-Risk. The performance of this Soft-Computing approach is analyzed using a set of assets from the Spanish stock market. Keywords: Evolutionary Multi-Objective Optimization, LR-Type Fuzzy Variables, MultiObjective Optimization, Portfolio Selection, Uncertainty Modelling. 196 FR3F (contributed) Fuzzy VIKOR Approach for Assessing the Vulnerability of the Water Supply to Climate Change and Variability in South Korea Chung, Eun-Sung (*) Seoul Nat University of Science & Tech Lee, Boram Seoul Nat University of Science & Tech Won, Kwang-Jae Seoul Nat University of Science & Tech This study assesses the vulnerability of the water supply to climate change and variability in the South Korean provinces for the present and future with a modified fuzzy VIKOR approach. Adaptive capacity and climate exposure, the key indicators of vulnerability, are profiled with the Delphi surveys based on the vulnerability concept, in which vulnerability is defined as a function of sensitivity. The fuzzy VIKOR method is used to aggregate the key indicators into a vulnerability score because it provides a compromise solution considering overall satisfaction and regret of the selection of wrong provinces. Markedly different rankings between the fuzzy VIKOR method and a conventional weighted sum method suggest the importance of considering regret of the selection of wrong provinces and including the uncertainty of information in the vulnerability assessment and consequent decision making for adaptation. Furthermore, diverse vulnerability rankings with the six different future scenarios suggest the need for robust decision making given the uncertainty of such diverse scenarios. Keywords: Fuzzy Sets, Fuzzy VIKOR, TOPSIS, Water Resources Management. 197 MO3G (invited) Generating an Optimal Set of Alternatives using Value-Focused Thinking Siebert, Johannes (*) University of Bayreuth Broadness and quality of identified alternatives are the basis of sustainable decisions. Bond et al. (2008) found out that decision makers are not aware of their objectives. We have conducted an empirical study which test setup was a straightforward extension of Bond et al. (2008) to analyze whether decision makers are aware of their alternatives. Our study confirms the hypothesis that decision makers are not aware of their alternatives, too. According to Keeney (1992), Value-focused Thinking enhances broadness and quality of alternatives identified by decision makers. We conducted a second empirical study which measured broadness and quality of alternatives identified by participants under different conditions. The results confirm the hypothesis that stimulation due to thinking about ones objectives leads to more and better alternatives. We also analyze the hypothesis that means are especially fruitful grounds to create alternatives. We have tested the impacts of different stimuli on the number, broadness and measurements of quality of the alternatives created. The analyzed stimuli are the use of means objectives, fundamental objectives, a master list of objectives, and a means-network. Keeney, R. L. (1992). Value-Focused Thinking a Path to Creative Decisionmaking. Bond, S. D., Carlson, K. A., & Keeney, R. L. (2008). Generating Objectives: Can Decision Makers Articulate What They Want? Management Science Vol. 54, No. 1, pp. 56-70. Keywords: Alternatives, Empirical Study, Objectives, Value Focus Thinking. 198 FR2E (invited) Generation of the Choquet Optimal Set of Multiobjective Combinatorial Optimization Problems Lust, Thibaut (*) LIP6 - CNRS Rolland, Antoine Universit´e Lyon 2 We study the generation of Choquet optimal solutions of multiobjective combinatorial optimization problems. Choquet optimal solutions are solutions that optimize a Choquet integral. The Choquet integral is used as an aggregation function, presenting different parameters, and allowing to take into account the interactions between the objectives. We develop a new property that characterizes the Choquet optimal solutions. From this property, a general method to easily generate these solutions is defined. We apply the method to the multiobjective knapsack problem. We show that Choquet optimal solutions that are not weighted sum optimal solutions represent only a small proportion of the Choquet optimal solutions and are located in a specific area of the objective space, but are much harder to compute than weighted sum optimal solutions. Keywords: Choquet Integral, Combinatorial Optimization, Multi-Objective Optimization. 199 MO2E (contributed) Goal Programming Approach to Workforce Scheduling Problem for a Restaurant C ¸ o¨l, G¨ ul¸cin (*) Osmangazi University Hasg¨ ul, Servet Osmangazi University Workforce is one of the most important source of an enterprise and effective use of it is an important issue. Creating a schedule that is compatible with the employee skills and preferences increases quality of the services. In this study employees have preferences on their days- off and work shifts. In addition, each employee has different skills. Using goal programming approach employees are assigned to daily shifts, days- off and tasks as much as possible to suitable to their preferences and skills. In addition gaps between the shifts and deviations from maximum and minimum working hours an employee must work during the scheduling period are minimized. The proposed model is applied in a restaurant for 1 and 2 months scheduling periods with heterogenous workload using prioritized goal programming approach. The results are compared with the schedule that is created by team leader. As a result effective schedules are obtained using goal programming method. Keywords: Goal Programming, Preferences, Workforce Scheduling. 200 MO2B (contributed) Goal Programming Model for Management Accounting: A New Typology McGillis, Sheila (*) Laurentian University Aouni, Belaid Laurentian University The environment within which accountants practice is becoming increasingly complex. This complexity is evident in all fields of accounting including management accounting, financial accounting and auditing. In these fields, accountants make decisions that involve consideration of multiple conflicting and incommensurable dimensions. Accountants need to ensure that they effectively consider these factors in arriving at their decisions concerning these complex problems. The Goal Programming (GP) Model was widely utilized for dealing with accounting decision-making situations. Different GP variants have been applied within the fields of management accounting, financial accounting and auditing. The aim of this paper is to propose a typology that will serve as a guideline to identify the most appropriate variant of GP to deal with specific accounting related decision-making. Keywords: Applications of MCDM, Goal Programming, Management Accounting And Auditing, Preference Information. 201 TU1A (plenary) Goal Programming: An Overview of Current Developments and an Application to Offshore Wind Farm Modelling Jones, Dylan (*) University of Portsmouth This seminar will present the current state-of-the-art of the goal programming technique. An overview of currently proposed variants and their linkage to underlying decision-maker philosophies will be given. The use of goal programming as a flexible modelling paradigm rather than a specific computational technique will be discussed. The combination and integration of goal programming with other modelling techniques from artificial intelligence and the wider field of multi-criteria decision making and Operational Research will be detailed. The use of meta-objectives to incorporate a mixture of underlying decisionmaker philosophies will be outlined. A modelling framework for this purpose will be developed. The second part of the seminar will be concerned with the application of Multi-Criteria Decision Making methodologies to the logistics problems arising during the life cycle of an offshore wind farm. The different stakeholders and their preferences will be outlined. Operational and strategic goals, objectives, and constraints will be formulated. The methodology outlined in the first section of the talk will be used in order to build a model that can be used to explore consensus solutions. The results will be discussed and conclusions drawn. Keywords: Goal Programming, Logistics, Offshore Wind Farm, State Of The Art. 202 TU3F (invited) GP-DEMO: Differential Evolution for Multiobjective Optimization Based on Gaussian Process Models Mlakar, Miha (*) Filipiˇc, Bogdan Petelin, Dejan Joˇzef Stefan Int. Postgraduate School Tuˇsar, Tea This paper proposes a novel surrogate-model-based evolutionary multiobjective algorithm called Differential Evolution for Multiobjective Optimization Based on Gaussian Process Models (GP-DEMO). The algorithm builds upon the newly defined relations for comparing solutions under uncertainty that extend the Pareto dominance relation. These newly defined relations, due to the use of a confidence interval of the approximation in addition to the approximated value, minimize the possibility of wrongly performed comparisons of solutions due to inaccurate surrogate model approximations. The GP-DEMO algorithm was tested on several benchmark problems and two computationally expensive real-world problems. To be able to assess the algorithm performance, we compared it with another surrogate-model-based algorithm called Generational Evolution Control (GEC), which does not consider uncertainty when comparing solutions, and Differential Evolution for Multiobjective Optimization (DEMO), which does not use surrogate models, but is known to be very effective. All three algorithms use the solution creation techniques from the Differential Evolution algorithm, and both GP-DEMO and GEC rely on Gaussian process modeling for building surrogate models. The DEMO algorithm was used for assessing the quality of the results gained with GP-DEMO, while the comparison between GP-DEMO and GEC was carried out to test how the new relations affect the number of exact solution evaluations needed during the optimization process. The analysis of the results over all test problems shows that the quality of the results obtained with GP-DEMO is similar to the results obtained with DEMO. The number of exactly evaluated solutions performed with GP-DEMO was similar to GEC, but the quality of the results obtained with GPDEMO was higher, mainly due to wrongly performed comparisons of the inaccurately approximated solutions in GEC. Keywords: Differential Evolution, Evolutionary Algorithm, Gaussian Process Modeling, Multi-Objective Optimization, Surrogate Modeling, Uncertain Pareto Dominance. 203 WE1E (invited) Green Economy in the State of Rio de Janeiro - A Multicriteria Evaluation Valle, Rogerio (*) Federal University of Rio de Janeiro Cl´ımaco, Jo˜ao INESC Coimbra-Universidade de Coimbra Rio de Janeiro commissioned a methodology for measuring the State’s Green Economy level. Economic, social and environmental aspects were prioritized and potential performance partial indicators were identified. It was decided to use data for each industry. Available databases provided our final set of partial indicators, which allowed measuring the multidimensional performance of each industry. It was decided not to aggregate the set of partial indicators into a single Green Economy index. For each partial indicator, thresholds were defined determining the tolerance range in the classification of industries performance. The aggregation among the dimensions, looking for a global classification of the industries, was done utilizing an interactive version of the conjunctive method. The SABILOC-MAT 1.0 package enables a clear communication among all the actors involved in the process. A dashboard was built enabling a simple and interactive sensitivity analysis. Keywords: Multi-Criteria Decision Analysis, Sustainability. 204 MO3C (contributed) Grey Numbers and Multi-Criteria Decision Making Pearman, Alan (*) University of Leeds Xie, Naiming Nanjing University of Aeronautics and Astronautics This paper is concerned with the standard Multi-Attribute Decision Making (MADM) model [m alternatives; n evaluation attributes, each with a weight, w; choice by maximising linear weighted sum of the scores]. The MADM model is typically an approximation. Approximation may stem from doubts whether all relevant attributes have been identified, about preference structures, presence of risk, and lack of confidence about accuracy of weights and/or scores. In this paper we focus on the last two of these. A standard approach in the face of such uncertainty has been to run the basic MADM model and then apply sensitivity testing. A second has been to recognise uncertainties explicitly from the outset, undertaking the analysis using devices to permit imprecise inputs, such as interval bounds on parameter values, or fuzzy numbers. An important recent innovation of this kind has been the development of grey analysis, following the introduction of grey systems theory (GST) in 1982 by Professor Deng. As the basic element of GST, grey numbers are defined as uncertain numbers such that the certain value is not known, although the potential value set is known. There are now a considerable number of papers outlining different grey approaches to multi-criteria decision making. In this paper we redefine the concept of grey number, outline the main features of grey analysis, review the range of grey multi-attribute choice models, identify strengths and weaknesses and compare the implementation of these grey models with others that have broadly similar aims. Then we create a new grey decision-making model by combining grey numbers with the MADM model. Assessment of the differences emphasises especially a behavioural focus, highlighting what sorts of judgement errors unaided decision makers might be most likely to make and what sorts of modelling support are best placed to help them avoid perceptual errors and have justifiable confidence in the choices they make. Keywords: Grey Numbers, Interval-Valued Functions, Multi-Criteria Decision Analysis. 205 TH2F (invited) Habitat Availability and Spatial Quality as Conflicting Objectives of Multi-use Forest Management Podkopaev, Dmitry (*) Mazziotta, Adriano Miettinen, Kaisa M¨onkk¨onen, Mikko Reunanen, Pasi University of Jyv¨askyl¨a Juutinen, Artti University of Oulu Tikkanen, Olli-Pekka University of Eastern Finland We consider a multiple-use forest management problem where goals are to maximize net present value of timber production and ecological quality of the forest landscape evaluated for several representative species. For each species we consider separately two indicators: habitat availability and landscape quality related to spatial arrangement of the habitat. This leads to formulation of a multi-choice knapsack problem with multiple linear and quadratic objectives. We address challenges of solving the problem for large forest landscapes and analyze trade-offs between habitat availability and quality of the habitat structure. Keywords: Combinatorial Optimization, Forest Management, Knapsack, Land Use Planning, Spatial Multi-Criteria Evaluation. 206 TH2E (invited) Handling Computational Variability in the Evaluation of Objectives: Ideas and Prospects Handl, Julia (*) The University of Manchester Knowles, Joshua The University of Manchester Allmendinger, Richard University College London Lavygina, Anna King’s College, London Our research focuses on multiobjective optimization scenarios in which the different objective functions have different computational or temporal costs, as may be the case e.g. in experimental optimization problems. Such scenarios may pose several challenges to algorithm design including (i) missing objective values, (ii) maintaining a set of solutions at once (e.g. to assess diversity or compute performance metrics) as more expensive objectives need longer to be evaluated, and (iii) integration of user preferences hampered by inaccurate objective values. Our previous work concluded that waiting for expensive evaluations to complete is the preferred choice when waiting periods are short, and whilst an EA augmented with a fitness-inheritance based approach for dealing with missing objective values performed well, we believe that other strategies may fare better, including surrogate-modelling approaches (e.g. Kriging) and strategies that schedule intelligently the evaluation of solutions on expensive objectives so as to maximize expected gain in information. Alternatively, EAs could be extended with reinforcement learning to learn when to switch between waiting and not waiting, and different strategies for dealing with missing values. To address (ii) one may allow mixing of generations so that solutions residing in an archive are getting progressively more accurately evaluated. Refining user preferences (interactively) as more expensive objective values become available can be an option to cope with (iii). To establish a common ground and facilitate algorithm design it seems to be worthwhile to define a general framework for describing algorithms that can deal with computationally variable objectives. This, combined with our empirical experience could form a good basis for more theoretical work on convergence and/or drift effects. on the simple algorithm models. Delayed constraints or resource constraints causing delays may also be considered in future. Keywords: Evolutionary Multi-Objective Optimization, Multi-Objective Optimization, Non-Regular Problems, Preferences. 207 TU5C (contributed) How to Support Environmental Management by MCDM Techniques? Reichert, Peter (*) Eawag Langhans, Simone Eawag Lienert, Judit Eawag Schuwirth, Nele Eawag Environmental decision support intends to use the best available scientific knowledge to help decision makers in finding management options that fulfill societal objectives to the degree possible. This requires a careful analysis of (i) how scientific knowledge can be formally represented and quantified, (ii) how societal preferences can be described and elicited, and (iii) how these theoretical concepts can best be used for communication with authorities, politicians, and the public. A multitude of MCDM methodologies is available to support these processes. We analyze and discuss which of these are conceptually most satisfying for this purpose and how practical limitations can be overcome. With respect to (i) we suggest the use of intersubjective probabilities to describe the current state of scientific knowledge. To address (ii), we suggest using value and expected utility theories to describe societal preferences. With respect to (iii) we outline a well-structured procedure for environmental decision support, exemplified for the case of river management. Key elements of this approach are the in-depth discussion, hierarchical structuring and quantification of objectives, the use of probabilistic scientific predictions, and the combination of these two elements for ranking and analyzing decision alternatives. This approach is complemented by visualization techniques to support transparent communication. We discuss conceptual arguments in favor of the suggested approach, its limitations, and first experiences of transferring it into river management practice in Switzerland. Keywords: Bayesian Analysis, Decision Support, Environmental Decision Making, MultiAttribute Utility Theory, Uncertainty Modelling. 208 MO5E (award) Hybrid Evolutionary Multi-Objective Optimization for Enhanced Convergence and Diversity Sindhya, Karthik (*) University of Jyv¨askyl¨a Evolutionary multi-objective optimization (EMO) algorithms are often criticized for their slow convergence, the lack of a theoretical convergence proof and for having no efficient termination criterion. In our research, the main focus is to improve EMO algorithms by addressing these criticisms. Hybrid EMO algorithms defined as hybrids of EMO algorithms and a local search procedure are proposed to overcome the criticisms of EMO algorithms. In the local search procedure, a local search operator originating from the field of MCDM (involving solving an achievement scalarizing function using an appropriate mathematical optimization technique) is used to enhance the convergence speed of a hybrid EMO algorithm. A hybrid framework, a base on which hybrid EMO algorithms can be built, is also proposed incorporating a local search procedure, an enhanced diversity preservation technique and a termination criterion. As a case study, a hybrid EMO algorithm based on the hybrid framework is successfully used to find Pareto optimal solutions desirable to a decision maker in the optimal control problem of a continuous casting of steel process. In addition, a hybrid mutation operator consisting of both nonlinear curve tracking and linear differential evolution mutation operators is proposed to handle various interdependencies between decision variables in an effective way. The efficacy of the hybrid operator is demonstrated with extensive numerical experiments on a number of test problems. Furthermore, a new progressively interactive evolutionary algorithm (PIE) is proposed to obtain a single solution desirable to the decision maker. Here an evolutionary algorithm is used to solve scalarizing functions formulated using the preference information of the decision maker. In PIE, the decision maker moves progressively towards her/his preferred solution by exploring and examining different solutions and does not have to trade-off between the objectives. Keywords: Achievement Scalarizing Functions, Differential Evolution, Hybrid Framework, Interactive EMO, Multi-Criteria Decision Making, Mutation, NAUTILUS Method, NSGA-II. 209 MO3D (contributed) Hydrological Constraints in Optimal Timber Harvest Scheduling Problems: A Case Study in Eucalyptus Plantations Gim´enez, Juan Carlos (*) Universidad de Extremadura Bertomeu Garc´ıa, Mercedes Universidad de Extremadura D´ıaz-Balteiro, Luis Universidad Polit´ecnica de Madrid This work describes a multiple criteria decision-making (MCDM) approach which forest managers can use to evaluate the potential impact of timber harvests on long-term forest management strategic objectives. We proposed a sequential procedure under a multicriteria framework to address sustainable management in industrial forest plantations taking into consideration hydrological constraints to reduce erosion risk. The approach is based upon a strategic forest-level planning model and Geographical Information System (GIS) data. First, a timber harvest scheduling model consistent with multiple objectives commonly identified in forest management plans for this type of plantations was formulated. Hydrological constraints were formulated to ensure erosion control for each stand. This formulation is based on GIS layers as slope, flow accumulation and LS factor (topographical USLE factor) obtained through spatial analysis. Then, a set of sustainability indicators was identified and briefly described. Next, several management alternatives were generated and assessed in terms of the sustainability indicators’ values. Finally, the management alternatives were ranked using a goal programming methodology by which the different indicators are properly aggregated. Furthermore, different priority rankings were obtained, which could represent potential preferences of the decision maker. The procedure was applied to a Eucalyptus globulus Labill. plantation in northwestern Spain. The results showed the validity of this approach, thus deriving the most sustainable management alternative under each scenario. Moreover, this methodology allows for consideration of different preferential weights for the sustainability indicators previously defined. In a further work, we plan to develop a tactical model to identify an annual silvicultural schedule map, i. e., we will try to precisely locate harvest areas in each stand under additional erosion control constraints. Keywords: Erosion Control, Eucalyptus Plantations, Goal Programming, Harvest Scheduling. 210 TU5F (contributed) Identifying Supported and Extreme Supported Nondominated Solutions and an Empirical Analysis of their Representative Quality Sayin, Serpil (*) Koc University Multiple Objective Discrete Optimization (MODO) problems are known to be burdened by the existence of unsupported efficient solutions. Previous work has demonstrated that supported efficient solutions, which are relatively easy to obtain, constitute only a small fraction of the efficient solutions, especially for larger problems. Obtaining the set of supported nondominated solutions has been a natural first step towards obtaining the entire set of nondominated solutions in some MODO approaches. To date, the quality of the set of supported nondominated solutions as a representation has not been thoroughly investigated. In this study, we investigate the representation quality of the set of supported nondominated solutions and the set of extreme supported nondominated solutions on the multiobjective knapsack and assignment problems. We obtain the entire efficient nondominated set using an epsilon-constraint-based approach. Given this set, we propose two simple linear programming formulations to identify the supported nondominated solutions and extreme supported nondominated solutions respectively. We experiment with different data generation schemes and different problem sizes. We observe that the set of supported nondominated solutions is almost always a good representation of the entire nondominated set. Our results indicate that even extreme supported nondominated solutions provide a good representation of the entire efficient set. We conclude that more research emphasis may be placed on obtaining supported nondominated solutions and extreme supported nondominated solutions due to the relative computational ease of obtaining them. Keywords: Multi-Objective Optimization, Non-Dominated Set, Representation Quality, Representative Set, Supported Non-Dominated Solution. 211 MO5H (poster) Impact of Food Price Volatility over Irrigation Agriculture using a Revealed Preference Model Gutierrez-Martin, Carlos (*) University of C´ordoba Berbel, Julio University of C´ordoba G´omez, Carlos Mario University of Alcal´a In this paper we show a simulation tool which uses mathematical programming methods to reveal the implicit multiattribute objective function lying behind the observed cropping decision. The model takes different criteria such as profit maximization, risk aversion, avoidance of management complexities and so forth into account. In order to determine the feasible combination of attributes of this objective function the model considers the production possibility frontier explicitly as depending on market prices, policy incentives, availability of production factors, water irrigation facilities, agronomic vocation and other constraints. Once calibrated the model becomes a powerful tool to assess the impact of different scenarios such as crop prices volatility, water prices modifications, irrigation technique substitution and so on. The objective of this document is to identify and quantify the effects of price volatility over the cropping patter and the water use. Other relevant variables such as agricultural income and employment are also considered. To do so we develop a revealed preferences model that uses basic microeconomic theory to calibrate and simulate farmers’ preferences. This kind of models present at least two advantages compared to the more commonly used Positive Mathematical Programming (PMP) (Heckelei and Britz, 2005; De Frahan et al., 2007) and Multi Criteria Decision Models (MCDM) (Berbel and Rodr´ıguez-Oca˜ na, 1998; Berbel and G´omez-Lim´on, 2000; G´omez-Lim´on and Riesgo, 2004): first, they do not assume linear preferences; and secondly, they do not assume implicit cost functions that are not observable, being both the main criticism against PMP and MCDM. The model is supported by a data base built on purpose for its implementation covering all Guadalquivir River Basin (Southern Spain) with high spatial detail, although we only presents the results of one of the typologies found in a cluster analysis. Keywords: Decision Maker’s Preferences, Price Volatility, Revealed Preferences. 212 TU5F (contributed) Improved Constraint Handling Technique for Multi-Objective Optimization Maheshwari, Vaibhav (*) Rangaiah, Gade Pandu Sharma, Shivom National University of Singapore Penalty function and feasibility approaches are the two popular constraint handling techniques for solving constrained MOO problems by stochastic global optimization methods. In general, feasibility approach works well for solving problems with inequality constraints. It gives higher priority to a feasible solution over an infeasible solution, but limits the diversity of search. Feasible search space is extremely small for equality constrained problems, and so feasibility approach may not be effective for solving problems with equality constraints. In this work, an adaptive relaxation of constraints along with feasibility approach is explored for solving constrained MOO problems by a stochastic optimizer. Here, all constraints are dynamically relaxed, which makes certain individuals temporarily feasible during selection of individuals for the next generation. The adaptive relaxation of constraints along with feasibility approach is incorporated in the multi-objective differential evolution algorithm, and its performance is compared with that of feasibility approach alone. It is first tested on two MOO test problems with equality constraints. Then, the proposed approach is applied to optimize the performance of two fermentation processes with many equality constraints. It is found that the proposed approach works better than the feasibility approach with a user defined relaxation value for equality constraints. However, both these simultaneous solution approaches perform inferior compared to sequential solution of optimization problem and equality constraints, on both the applications. Sequential approach requires numerical techniques for solving highly complex and nonlinear process model equations, and these techniques may require good initial estimates. Conversely, simultaneous solution approach does not require any additional information. In this work, the proposed approach will be described, and will be compared with the feasibility and sequential approaches. Keywords: Adaptive Relaxation Of Constraints, Constraint Handling, Feasibility Approach, Multi-Objective Differential Evolution, Sequential Solution Approach, Simultaneous Solution Approach. 213 TH2B (contributed) Improved Upper Bounds for a Two-Phase Biobjective Shortest Path Algorithm Medrano, F. Antonio (*) Church, Richard University of California Santa Barbara The biobjective shortest path (BOSP) problem is an extension of the single-objective shortest path problem, in that it simultaneous considers optimizing two competing objectives. Rather than solving for a single optimum, the BOSP solves a set of paretooptimal paths that represent the best trade-offs between the two objectives. This has been shown to be NP-hard. Raith and Ehrgott (2009) compared the existing state-of-theart approaches for solving BOSP problems, including label-correcting and label-setting methods, and a new enumeration method using a Near Shortest Path (NSP) algorithm from Carlyle and Wood (2005). They also proposed two-phase variants of the three approaches that use a polynomial technique for determining supported solutions to define an upper-bound on the unsupported solution search. We have developed a new algorithm based on the two-phase NSP method proposed by Raith and Ehrgott. Our new method sets a tighter initial upper bound by generating gateway paths (Lombard and Church 1993) during the supported solution computation (Cohon et al. 1979). With minimal additional computation, the gateway heuristic generates a set of paths that are near-optimal to the paretal set. This tighter upper bound reduces the NSP computation, resulting in faster overall run-times. We discuss the details of this approach, and results comparing it to other methods on various networks. Carlyle, W.M. & R.K. Wood, (2005). Nearshortest and k-shortest simple paths. Networks, 46, 98-109. Cohon, J.L., R.L. Church & D.P. Sheer, (1979). Generating multiobjective trade-offs: An algorithm for bicriterion problems. Water Resources Research, 15, 1001-1010. Lombard, K. & R. Church, (1993). The gateway shortest path problem: Generating alternative routes for a corridor location problem. Geographical Systems, 1, 25-45. Raith, A. & M. Ehrgott, (2009). A comparison of solution strategies for biobjective shortest path problems. Computers & Operations Research, 36, 1299-1331. Keywords: Corridor Location, Multi-Objective Optimization, Multi-Objective Shortest Path Problem, Shortest Path. 214 TU3G (contributed) Incorporating Stakeholders’ Preferences into Wastewater Infrastructure Planning in Switzerland Zheng, Jun (*) Eawag Lienert, Judit Eawag We consider a complex decision problem aiming at planning wastewater infrastructures until 2050 in Switzerland. The planning has long-term environmental, economic and social consequences, as assets are expensive and long-lived. Incorporating preferences of stakeholders representing different social sectors into the decision process enables them to get an integrated perspective and ultimately helps to reach a consensus. Eliciting stakeholders’ preferences with Multi-attribute utility theory (MAUT) can be cognitively demanding for the stakeholders whose ambiguous value judgments are to be quantified. We tackle the difficulty by (1) using adapted interval SWING method which elicits weights of objectives in an imprecise way; (2) decomposing weight elicitation questions into four types of simple questions, which significantly reduce their cognitive burden so that the weights are obtained in a reliable way; (3) allowing stakeholders to express preferences with verbal scales that are later translated to numbers, rather than directly asking for numbers; (4) using structured and personalized questions to elicit value functions and risk attitudes of their most important objectives, and their attitudes for tradeoffs which determine their own aggregation models. The questions are designed based on behavioral decision theory to overcome common biases. They are implemented by an online questionnaire and face-to-face structured interviews with 10 important stakeholders. We reveal different stakeholders’ preferences towards the objectives of wastewater infrastructure planning. Preference programming technique then identifies the most robust alternative as a “best” alternative for all stakeholders using the elicited imprecise preferences. The proposed elicitation procedure is applied in the context of wastewater infrastructure planning, but is widely applicable when the stakeholders’ preferences for complex decision problems are to be incorporated into MAUT framework. Keywords: Environmental Decision Making, Preferences, Utility Theory, Water Infrastructure Planning, Weights. 215 TU4C (contributed) Incorporation and Communication of Uncertainty in Decision Support for Environmental Management: A Case Study on Water Quality Improvement Schuwirth, Nele (*) Eawag Reichert, Peter Eawag Stamm, Christian Eawag We illustrate methods to incorporate uncertainty into environmental decision support processes with a case study on the improvement of water quality in a river catchment of the Swiss Plateau. Different management options to reduce point and non-point sources are evaluated in this study. We apply the multi-attribute value and utility theory to evaluate alternatives and demonstrate how methods of sensitivity and uncertainty analysis can be used to facilitate the decision making process. Uncertainty in prediction of consequences formally requires the consideration of risk attitude in addition to the strength of preference and therefore switching from value to utility functions. However, the elicitation of utility instead of value functions has major disadvantages and is much more demanding. In complex decision situations with many sub-objectives this might not be feasible. We examine under which conditions risk attitude affects the ranking of alternatives and when the information contained in value functions is sufficient and even superior for decision support. Along with the case study, we introduce the software package ’utility’ (implemented in the statistics and graphics environment R) that facilitates the visualization of uncertain valuations in complex decision situations. The package allows its users to define arbitrary aggregation functions, to formulate multi-attribute and conditional end nodes, to switch from values to utilities at any level of the objectives hierarchy, and to extend the decision support framework by arbitrary types of nodes. With these features it is intended to facilitate research on the construction of value and utility functions beyond the additive model that appears to be inappropriate in certain environmental management situations. Keywords: Environmental Decision Making, Multi-Criteria Decision Analysis, Risk, Uncertainty Modelling, Utility Theory. 216 TH3C (contributed) Increasing the Transparency of Model Selection in Multi-Criterion Data-Clustering Handl, Julia (*) The University of Manchester Papamichail, K.Nadia The University of Manchester State-of-the-art systems for decision-support typically include functions to explain and to justify the decisions recommended by the system (Papamichail & French 2003). Capabilities of this type have been shown to play an essential role in informing decision makers and in ensuring the adoption of the recommendations made by the system. In this work, we consider the potential use of similar approaches in the context of multicriterion methods for data clustering (Handl & Knowles 2007). In particular, we are looking at the automatic generation of contextual explanations and justifications using natural language generation techniques to support the comparison and selection of representative solutions in multi-criterion clustering, and increase the transparency of existing approaches. Methods for data-clustering use intrinsic data properties to partition large data sets into clusters, with the ultimate aim of identifying groupings that are meaningful and interpretable by an expert user. Strategies to support decision makers in this last step have received little attention in the literature. Generally, no attempt has been made to support contextual comparisons between sets of different possible clustering solutions. This is despite the fact that sets of fundamentally different solutions are commonly encountered as the end result of a cluster analysis: this may be as a consequence of the use of various clustering algorithms, different parameter settings, or due to multiple clustering objectives. Solutions within the final set of candidate partitionings will typically differ with respect to the number of clusters, the level of discrimination along individual features, the scores for specific clustering criteria and their mutual similarity. Our aim is to develop and present a system that integrates such information and facilitates contextual comparisons between solutions. This would enable experts to make more informed decisions about their preferred clustering solutions Keywords: Expert systems, Multi-Criteria Decision Analysis. 217 MO5H (poster) Induction of Belief Rules and Belief Decision Trees from Uncertain Data AbuDahab, Khalil (*) The University of Manchester Chen, Yu-Wang The University of Manchester Xu, Ling The University of Manchester In this paper, we present and evaluate an induction algorithm for acquiring belief rulebases, by inductive inference, from uncertain data. Existing methods are able to extract traditional rules (with consequents that are believed to be either 100% true or 100% false) inductively from certain data. The proposed method extracts belief rules from data which may contain uncertain or incomplete knowledge in both antecedent and consequent attributes. Belief rules can capture uncertain or incomplete knowledge using uncertain belief degrees in consequents. Instead of using singled-value consequents, each belief rule deals with a set of collectively exhaustive and mutually exclusive consequents. The method is also able to handle hierarchically structured attributes. A MCDM aggregation method is used to aggregate attributes then only the aggregated attributes are included in the rules and used during inference process. A number of aggregation methods have been evaluated to find the most appropriate one for calculating an aggregated information entropy. The evidential reasoning (ER) approach was able to handle multiple, qualitative or quantitative, attributes under uncertainty while other methods fail to do so. These features allow to minimise tree size in order to avoid over-fitting the training data and poorly generalising to new samples. They also avoid producing a larger number of rules which can make the rule-based reasoning system slow. A generic expert system was developed in order to evaluate the algorithm. The system allows rule-based classification models to be created by acquiring knowledge inductively from training examples or deductively from experts in the problem domain. Building belief rule based models inductively from data and deductively from domain knowledge is important when domain experts have sufficient knowledge of some, but not all, aspects of a problem. Keywords: Belief Decision Trees, Classification, Rule Extraction, Rule-Based Knowledge Acquisition, Supervised Learning. 218 TH3B (invited) Inference of Parsimonious Preference Models in Robust Ordinal Regression Slowinski, Roman (*) Poznan University of Technology Greco, Salvatore University of Catania Mousseau, Vincent Ecole Centrale Paris Ordinal regression is used to infer parameters of a supposed preference model, which make this model compatible with preference information provided by the Decision Maker (DM). As values of these parameters are usually not unique, robust ordinal regression handles a complete set of parameters of the compatible preference model of a given class (e.g., additive value function, like in UTAˆGMS method, or outranking relation, like in ELECTREˆGKMS or PROMETHEEˆGKS methods). Here we care not only about compatibility of the model with the preference information, but also about keeping the model as parsimonious as possible. In this context, “parsimonious” means simple and intuitive. We achieve this goal by solving two linear programming problems that permit to find the “most linear” value function able to represent the considered preferences, as well as the value function composed of the minimal number of linear pieces, respectively. These two linear programming problems have a solution under condition that it is possible to represent the DM’s preference information using an additive value function. If this is not the case, some more complex value function can be considered, for example, involving synergy components for pairs of criteria, as in the UTAˆGMS–INT method. Also in this case, one can apply the above approach in order to search for a parsimonious model, i.e. “most linear” marginal value functions with the least number of pairwise synergy components. Keywords: Elicitation, Interaction Among Criteria, Multi-Criteria Decision Aid, Preference Learning, Preference Modeling, Robust Ordinal Regression. 219 TU2E (invited) Information Levels in Additive Group Decision Models under Incomplete Information: Bridging the Cardinal-Ordinal Gap Dias, Luis (*) University Coimbra Sarabando, Paula IP Viseu - INESC Coimbra Vetschera, Rudolf University of Vienna Aggregation of cardinal values in groups requires group members to specify their preferences in cardinal form, by assigning utility scores to alternatives. This could be difficult for group members, in particular if this information should be provided exactly. Methods for decision making under incomplete information like robust ordinal regression could be employed to simplify this task. Such methods allow information to be specified also only in ordinal form (e.g. as a ranking); this information can then be used to approximate a cardinal utility function. Along the spectrum between purely cardinal and purely ordinal information, group members could also specify preference information not only as a ranking of alternatives, but also as a ranking of differences of alternatives. Obviously, providing information closer to exact cardinal utility values will increase the precision of preference statements at the group level, but the strength of this effect is not clear. Besides each individual member’s incompletely known preference function, the weighting of these functions is also unknown. A recent paper extended Arrow’s non-dictatorship condition to the aggregation of cardinal preferences in a way that implies a constrained set of weights of group member in an additive group utility function. In this paper, we report on a comprehensive computational study, in which we attempted to quantify the effect of providing different levels of preference information (exact cardinal values, rankings of differences between alternatives and rankings of alternatives), and of the imposition of different constraints on members’ weights in an additive group utility function (equal weights, general random weights and random weights satisfying the Non-dictatorship condition) on outcomes at the group level. Relevant outcome dimensions analyzed refer to the structure of the group preference relation (necessary and possible relations) and the strength of impact of individual group members. Keywords: Group Decision Making, Incomplete Information, Robust Ordinal Regression. 220 MO4E (contributed) Integrating Qualitative Assessment and Fuzzy AHP to Prioritize Investments in an Energy Efficiency Program of Favelas in the City of Rio de Janeiro Ribas, Jose Roberto (*) Universidade Federal do Rio de Janeiro Rocha, Mariana Light Servi¸cos de Eletricidade S/A Severo, Juliana Ribas Delta Energia This study proposes an integrated qualitative and quantitative assessment of expert opinions aiming at rank a set of five favelas located in the city of Rio de Janeiro candidates for investments in an energy efficiency program implemented by the local electric utility company. The city wants to get rid of crime and present itself as a peaceful metropolis to host two big sports events: the soccer’s World Cup in 2014 and the Olympic Games in 2016. In the past years some favelas were chosen to be prototype for an ambitious project to reshape Rio de Janeiro, known by its acronym, UPP (Unidade de Policia Pacificadora), which means Pacifying Police Unit. This involves sending in special tactics police (BOPE) to drive drug traffickers out, then installing the UPP in the favela. Once security has been established is possible to improve general living conditions in these areas. The core of the energy efficiency program was to convert informal customers to formal ones, as far as such communities were responsible for approximately 40% of the commercial loss (stolen energy) in the city. The model specification presented in this paper was set up with ten relevant criteria for decision making identified through an in-depth interview with managers. The relative importance of the criteria and the performance of each favela regarding every criterion were measured by SIMOS method. The preferences resulting from this method were translated to a nine-point scale. The imprecision of subjective judgment was partially compensated by using a fuzzy analytical hierarchy process (FAHP). Some criteria were ordinal, such as “Fair Relationship with the Community” and “Complexity to Rebuild the Distribution Lines”, other were cardinal, like “Percentage of Clients in Default” or “Commercial Loss Due to Energy Theft”. At the end, the model was efficient on ranking the five favelas, therefore, contributing to a rational and transparent approach for capital investment in social projects. Keywords: Energy Efficiency, FAHP, Favelas, Prioritizing Objectives. 221 FR2C (invited) Learning the Parameters of a Multiple Criteria Sorting Method from Large Sets of Assignment Examples Sobrie, Olivier (*) Universit´e de Mons Mousseau, Vincent Ecole Centrale Paris Pirlot, Marc Universit´e de Mons In multiple criteria decision analysis (MCDA), the sorting procedures aim at assigning each alternative of a set into one of the pre-defined and ordered categories. We consider in this study a derived version of the original ELECTRE TRI sorting procedure. This simplified version of the ELECTRE TRI procedure, proposed and studied by Bouyssou and Marchant, assigns the alternatives into categories on the basis of a majority rule. We call it MR-Sort procedure. A difficulty with the MR-Sort procedure is the elicitation of its parameters. An MR-Sort model composed of n criteria and p category requires the elicitation of p - 1 profiles, n weights and a cut threshold. For the decision maker, eliciting explicitly these parameters is not trivial. This is why several papers deal with the learning of parameters of such models on the basis of assignment examples. Currently, the proposed algorithms are not suitable for large problems, i.e. problems involving a lot of learning examples, categories and/or criteria. This is mainly due to the computing time that becomes important when the number of parameters increases. We propose a metaheuristic to learn the parameters of a MR-Sort from a set of assignment examples and their performances vectors. The metaheuristic is composed of two main parts. The first part aims at learning the weights and the cut threshold and the second one aims at learning the profiles. The algorithm first generates a random set of profiles. Then a linear program is used to learn the weights and the cut threshold on basis of the random generated profiles. After learning the weights, a metaheuristic adjusts the profiles evaluations. The metaheuristic tries to maximize the number of compatible examples with the learned model. Our research includes a set of experiments studying the algorithm performances, the behavior of the metaheuristic in presence of errors in the assignment examples and the algorithm capability to find models restoring the input examples. Keywords: ELECTRE Methods, Linear Programming, Metaheuristics, Multi-Criteria Sorting, Preference Learning. 222 WE2G (invited) Location Decision for Firm Expansion: A Bicriterion Approach Pelegr´ın, Blas (*) Universidad de Murcia Fern´andez Hern´andez, Pascual Universidad de Murcia Garc´ıa P´erez, Mar´ıa Dolores Universudad Cat´olica San Antonio Location is an strategic decision for an expanding firm in competition with other existing firms. We present different location models in which two objectives are considered. The first is maximizing the market share (or profit) obtained by the expanding firm, once its new facilities are located. The second is minimizing the lost of market share (or profit) of its old facilities caused by the new facilities (called the cannibalization effect). Customers chose a facility according to a patronizing behaviour depending of each location model. Some formulations are presented which let generate Pareto solutions by solving integer mixed linear programming problems. Keywords: Decision Analysis, Location, Mixed Integer Programming. 223 MO2G (contributed) Markowitz’s Investment Problem under Multicriteria, Uncertainty and Risk Korotkov, Vladimir (*) Belarusian State University Emelichev, Vladimir Belarusian State University We consider a multicriteria Boolean discrete variant of Markowitz’s investment problem with Savage’s bottleneck risk criteria. Such model demonstrates the way how an investor, choosing the right portfolio, can minimize the risk of losing own gain, given some certain level of expected income. Under such formulation, the initial data are given as static expert evaluations of risks (financial, environmental and similar). It is well-known that calculating these values is usually accompanied by a large number of computational errors which lead to a high degree of uncertainty of the initial data. Thus, there is a need to take into account uncertainty and inaccuracy of the initial data for the investment problems like this, because even small changes of initial data make a model behave in an unpredictable manner. One of the crucial questions arising in uncertainty analysis is the question about extreme level of the initial data variations (perturbations) preserving optimality of a chosen portfolio. We investigate quantitative estimates of permissible changes of the uncertain initial date, which preserve the Pareto optimality of a portfolio using the H¨older metric in the perturbed parameters space. As a result we present attainable (unimprovable) lower and upper bounds for the stability radius. Keywords: H¨older Metric, Multi-Criteria Investment Problem, Pareto Optimal Investment Portfolio, Savage’s Criteria, Stability Radius Of Portfolio. 224 WE1E (contributed) MCDA-GIS Integration for the Evaluation of the Invasive Alien Species. The Case of the Eastern Grey Squirrel in Umbria (Central Italy) Rocchi, Lucia (*) Massei, Gianluca Paoloni, Daniele University of Perugia Multicriteria Decision Analysis (MCDA) methods are a basic tool in the field of environmental valuation, but they cannot take into account the spatial dimension easily. The conventional MCDA approach assumes spatial homogeneity of alternatives within the case study area, although this is unrealistic. To face a spatial decision problem means to combine Geographical Information Systems (GIS) and MCDA approach [2]. Value judgments, geographical data, their transformation and elaboration are needed to solve it[2]. Several integrations are possible, but only with a complete one [2] the two systems share the same interface and database. An example of complete integration is the r.mcda suite [3], we used in this work. The suite is a modular MCDA package developed in GRASS GIS 6.4, including several multicriteria methods. We applied the DRSA [1], for understanding the possible damage to agriculture crops by the grey squirrel. The Eastern grey squirrel (Sciurus carolinensis) is a North American species that was introduced in UK and Italy. Its impact on European ecosystems is very strong, in particular on the native red squirrel. Grey squirrels are able to compete more successfully than the reds ones for food and habitat. This competitive exclusion leads to the extinction of the native species [5]. Grey squirrels can cause damages also to agroecosystems, with heavy consequences for agriculture [4]. The aim of the work is to identify risk areas for high quality crops, related to the potential spreading of grey squirrel in Umbria (Italy). Several hotbeds have been identified, and without proper management and control actions by public bodies, the matter of Grey squirrel is destinated to blow up. [1]GRECO S., MATARAZZO B., SLOWINSKI R. 2001. Rough sets theory for multicriteria decision analysis. EUR J OPER RES 129, 1-47. [2]MALCZEWSKI J. 2006. Gis- based multicriteria decision analysis: a survey of the literature. IJGIS 20, 703-726. [3]MASSEI G, ROCCHI L., PAOLOTTI L., GRECO S., BOGGIA A. 2012.MCDA GIS integration: an application in GRASS GIS 6.4. In OGRS- Symposium proceeding O. Ertz, S. Joost and Tonini M., (Eds) (2012) pp 195-201. [4]SIGNORILE A.L., EVANS J. 2012 Damage caused by the American grey squirrel (Sciurus carolinensis) to agricultural crops, poplar plantations and semi-natural woodland in Piedmont, Italy. Forestry, 80, 89-98. [5]Wauters, L.A., Tosi, G., Gurnell, J., 2002. Interspecific competition in tree squirrels: do introduced gray squirrels (Sciurus carolinensis) deplete tree seeds hoarded by red squirrels (Sciurus vulgaris)? BEHAV ECOL SOCIOBIOL 51, 360-367. Keywords: Decision Support Systems, Dominance-Based Rough Set Approach , GIS, GISMCDA Integration, Invasive Alien Species. 225 MO2D (contributed) MCDM and GIS to Identify Land Suitability for Agriculture Abdelkader, Mendas (*) CTS The integration of MultiCriteria Decision Making approaches (MCDM) in a Geographical Information System (GIS) provides a powerful spatial decision support system which offers the opportunity to efficiently produce suitability maps. Indeed, GIS is a powerful tool for analyzing spatial data and establishing a process for decision support. Because of their spatial aggregation functions, MCDM methods can facilitate decision making in situations where several solutions are available, various criteria have to be taken into account and decision-makers are in conflict. A spatial decision support system has been developed for establishing the land suitability map for agriculture. It incorporates the multicriteria analysis method ELECTRE Tri (ELimitation Et Choix Traduisant la REalit´e) in a GIS within the GIS program package environment. The parameters and the classification system used in this work are inspired from the FAO (Food and Agriculture Organization) approach dedicated to a sustainable agriculture.The main purpose of this research is to propose a conceptual and methodological framework for the combination of GIS and multicriteria methods in a single coherent system that takes into account the whole process from the acquisition of spatially referenced data to decision-making. In this context, a spatial decision support system for developing land suitability maps for agriculture has been developed. The algorithm of ELECTRE Tri is incorporated into a GIS environment and added to the other analysis functions of GIS. This approach has been tested on an area in Algeria. A land suitability map for durum wheat has been produced. Through the obtained results, it appears that ELECTRE Tri method, integrated into a GIS, is better suited to the problem of land suitability for agriculture. The coherence of the obtained maps confirms the system effectiveness. Keywords: Assignment, Decision Analysis, Decision Support Systems, Multi-Criteria Decision Analysis, Preference Aggregation. 226 TH4D (contributed) MCDM in Nutritional Epidemiology Gerdessen, Johanna (*) Wageningen University Banasik, Aleksander Wageningen University Claassen, G.D.H. (Frits) Wageningen University The intended conference presentation aims to make the audience familiar with an MCDM challenge in nutritional epidemiology, an area so far hardly discovered by MCDM (and vice versa). Epidemiological studies investigate the relationship between diet and disease. Their relevance can hardly be overestimated, e.g. in designing effective and efficient intervention studies related to obesity, diabetes or cancer. For example, the hypothesis that consumption of red and processed meat increases colorectal cancer risk while intake of fish decreases risk is strongly supported by the results of an epidemiological study. Large epidemiological studies commonly use food frequency questionnaires (FFQs) to assess a population’s habitual (nutrient) intake. FFQs ask respondents about their consumption of several food items during a predefined time period. On the one hand an FFQ should include enough questions on food items to capture sufficient information on all nutrients of interest. On the other hand an FFQ should be as short as possible, because long FFQs are less cost and time efficient, may bore respondents and make them less motivated to fill out an FFQ accurately. Selection of questions on food items to be included in an FFQ is based on their contribution to the nutrient intake of a population. Aim of the selection procedure is to compose a set of food items that captures as much information as possible on a set of multiple nutrients. Each nutrient is represented with an objective function, which makes it a MCDM problem. The aim of this presentation is to show how MCDM can support the decision maker in the selection procedure. In consultation with the program committee the presentation can focus either on a general form of the model, or on the reformulation approach that was designed to solve a specific fractional programming instance with more than 200 fractional terms. Keywords: Healthcare, Linear Programming, Mixed Integer Programming. 227 WE2B (contributed) Measuring Attractiveness of High School Programs using Composite Index of Attractiveness and a Multi-level Outranking Framework: Evidence from PISA Survey in Serbia Stamenkovic, Mladen (*) University of Belgrade Anic, Ivan University of Belgrade Backovic, Marko University of Belgrade Popovic, Zoran University of Belgrade In this paper we research attractiveness of high school programs in Serbia for students with high and low level of competencies in three subjects: mathematics, reading and science. We use PISA 2009 data. PISA survey is held every three years by OECD since 1997. We split students in four different groups by their socio-economic status (ESCS). This status is given by ESCS index measured in PISA survey and conveys information about social, economic, cultural and educational status of student’s family. As measure of school attractiveness we create index of school attractiveness (ISA). We further investigate the attractiveness of high school programs by using ELECTRE MLO (ELECTRE Multi-Level Outranking) approach, well used for outranking alternatives, based on one modification of ELECTRE I method. The original ELECTRE I method is not good for hierarchical comparison on set of alternatives because of cycles that can possibly occur. In modification which we use in this paper cycles can be avoided by introducing the Absolute Significance Threshold. If the threshold value for the Concordance index is above the Absolute Significance Threshold value, than there will be no cycles in the relation graph. Moreover, the relation between alternatives is strictly ordered relation. Unlike ELECTRE I where we have alternatives considered as in core or outside the core, ELECTRE MLO calculates hierarchical levels of performance. The aim of the proposed ELECTRE MLO approach is to produce a specific grouping of the selected high school programs, based on the dominance relationships between them. We show that Gymnasiums have the most attractive school program in Serbia. This means that the biggest part of high skilled students and the lowest number of low skilled students goes to this program. Furthermore, the result we got using ISA and by ELECTRE MLO approach are in line and lead us to the same hierarchical structure of high school programs. Keywords: ELECTRE Methods, Outranking Methods, PISA Survey. 228 TU3C (contributed) Measuring Performance of European Countries in the Cultural Domain: an MCDM Approach Nisel, Seyhan (*) Istanbul University Nisel, Rauf Marmara University Recent initiatives and studies have been carried out by the European Commission in the field of culture have revealed the strategic importance of cultural aspects in sustainable development. The European commission has generated wide variety of statistical data for European countries in the field of culture, which provides the opportunity for extensive analysis. The aim of this study is to propose a performance measurement model for EU-27, EFTA and candidate countries in terms of cultural domain and practices. In the study, the scope of the cultural field, evaluating criteria and data for countries were obtained from cultural statistics that was published by Eurostat in 2011. The VIKOR (VlseKriterijumska Optimizacija I Kompromisno Resenje) methodology, applicable to many conflicting and incommensurable criteria, was used for calculating the overall performance of the countries. The research results showed the rank of the countries in terms of their performances in many different the cultural aspects. Keywords: Culture, EU, Performance Measurement, VIKOR. 229 MO3B (contributed) Microsoft Excel as a Tool for Solving Multicriteria Decision Problems Perzina, Radomir (*) Silesian University Ramik, Jaroslav Silesian University The Analytic Hierarchy Process is frequently used approach for solving decision making problems. There exists wide range of software programs utilizing that approach. Main disadvantage of those software products is that they are commercial and relatively expensive and thus it prevents them to be used by small companies, students or researchers. This contribution introduces a Microsoft Excel add-in called DAME – Decision Analysis Module for Excel. Comparing to other computer programs DAME is free, can work with scenarios or multiple decision makers, allows for easy manipulation with data and utilizes capabilities of Microsoft Excel. Users can structure their decision models into three levels – scenarios/users, criteria and variants. Items on all three levels can be evaluated either by weights or pair-wise comparisons. For each pair-wise comparison matrix there is calculated an inconsistency index. There are provided three different methods for the evaluation of the weights of criteria, the variants as well as the scenarios – Saaty’s Method, Geometric Mean Method and Fuller’s Triangle Method. Multiplicative and additive syntheses are supported. All calculations are instant so users can easily see what happen if anything is modified. Bar chart is used for final ordering representation. The proposed software package will be demonstrated on couple of illustrating examples of real life decision problems. Keywords: Analytic Hierarchy Process, Microsoft Excel, Multi-Criteria Decision Making, Pairwise Comparison, Scenarios, Software. 230 TU5D (contributed) Minmax Robustness for Multi-Objective Optimization Problems Ide, Jonas (*) University of G¨ottingen Ehrgott, Matthias Lancaster University Sch¨obel, Anita Georg-August University G¨ottingen In real-world applications of optimization, the best solution is not always helpful if disturbances or other changes to the input data occur. Such changes in fact occur frequently and can result in a non-optimality or even infeasibility of the optimal solution. One way to deal with uncertain input data is robust optimization, the aim of which is to find solutions which remain feasible and of good quality in all scenarios. For single-objective optimization problems several definitions of robustness, i.e. when a solution is seen as robust against such uncertainties, have been analyzed thoroughly. One of these concepts is the concept of minmax robustness, where a solution is called robust if it is feasible for every scenario and minimizes the objective function in the worst case. In this talk we show that by solving a set-valued maximization problem, the concept of minmax robustness can be extended to multi-objective (MO) optimization and call this extension minmax robust efficiency. We investigate properties of this concept, introduce strict and weak minmax robust efficiency and discuss how minmax robust efficient solutions can be computed. For the latter, we extend common techniques for calculating efficient solutions for deterministic MO problems such as the weighted sum scalarization and the epsilonconstraint method. We investigate which theoretical results of the original techniques can be adjusted to hold also in the uncertain setting and what relations between the techniques exist. As an extension we present other concepts of robustness which rely on other definitions of the set-valued maximum. We conclude the talk by pointing out possible areas of application such as portfolio optimization and supply chain management in the renewable resource sector. For the latter we present an example from the veneer-cutting industry where the cutoff and the storage volume have to be minimized with regard to the uncertain quality of the used wood. Keywords: Multi-Criteria Decision Making, Multi-Objective Optimization, Portfolio Optimization, Robust Supply Chains, Robustness And Sensitivity Analysis, Scenarios. 231 TU3D (contributed) Mixed Generation Sensitivity Analysis: An Optimal Scenario Study Wilson, Reginald (*) Seattle City Light There are several generating units available to meet the load demand on any given day. The mixed-integer program case engages generating units that can be shut down or operated between minimum and maximum output levels within a specific time demand block. An examination of these conditions espouses the utilization of renewable energy technology as a reliable power generation source as well as its effectiveness and augmentation in a robust, heavily subjugated thermal-generation system. The mixed integer problem has the predisposition to generate many sub-problems during the computation. However, with careful selection of the parameters in the model, the challenges are easily overcome. Hydroelectric energy is significant from an operational standpoint as it requires little or no ramp-up time, commonly found in many combustion technologies. The essential load-following capability, peaking capacity and voltage stability attributes renders hydroelectric energy as a significant part in ensuring reliable electricity service and in meeting customer needs in a market driven industry. When this renewable resource is combined with the thermal generation, an improved reliability function is apparent. Yet, in order to minimize the environmental effects inherent in this type of energy resource, the renewable power supply is expected to produce an effectual, consistent output. The integration of renewable generation within the electrical system entails intricate analysis of the transmission line upgrades throughout the system. These upgrades are designed to relieve congestion, improve reliability, and provide voltage support by thoroughly evaluating the alternatives to accommodate the new expansion efforts. This paper examines the system integration of renewable generators, via simulated studies, in the context of electrical operations from the hydroelectric generation perspective. Keywords: Cost Avoidance, Mixed Integer Programming, Power Generation, Production Cost, Renewable Energy, Scenarios, Sensitivity Analysis, Smart Devices, System Integration, Trade-Offs. 232 WE2E (contributed) Mixed Integer Multi-Objective Optimization for Flight and Maintenance Planning of Mission Aircraft Gavranis, Andreas (*) University of Thessaly Kozanidis, George University of Thessaly Flight and Maintenance Planning (FMP) of mission aircraft addresses the question of which available aircraft to fly and for how long, and which grounded aircraft to perform maintenance operations on, in a group of aircraft that comprise a unit. The objective is to achieve maximum fleet availability of the unit over a given planning horizon, while also satisfying certain flight and maintenance requirements. Solution approaches that have previously been developed to tackle this problem often perform unsatisfactorily, providing a long-term fleet availability with significant variability (defined as the sum of squares of the deviations of the individual fleet availability values from their corresponding average value). In order to address this issue, we introduce an additional objective that minimizes the variability of the fleet availability, and we develop an exact solution algorithm for the resulting multi-objective FMP model that is capable of identifying the entire frontier of non-dominated solutions. The algorithm utilizes the discrete nature of the domain of the fleet availability values, which is due to the fact that the fleet availability depends solely on the number of different combinations of aircraft that can enter and exit the maintenance station. Initially, the algorithm identifies an upper bound on the optimal fleet availability by solving a simplified relaxation of the original problem; then this value is gradually increased, and feasible aircraft combinations that minimize the variability of the associated fleet availability are identified. Employing special valid inequalities for excluding solutions that do not qualify for optimality, the algorithm obtains the entire frontier of non-dominated solutions. The experimental results that we present demonstrate its high efficiency on randomly generated and realistic problem instances. Keywords: Fleet Availability, Flight And Maintenance Planning, Mixed Integer MultiObjective Programming, Non-Dominated Frontier. 233 FR2D (contributed) Modeling Consumer Preferences about Vehicles with Multi-Attribute Additive Models: Survey-Based Experiments Dias, Luis (*) University Coimbra Oliveira, Gabriela INESC Coimbra and University of Coimbra - MIT Portugal Sarabando, Paula IP Viseu - INESC Coimbra This study concerns the choice of a vehicle focusing on its powertrain technology: gasoline, diesel, hybrid, plug-in hybrid or fully electric. The purchase of a vehicle is a standard example of a problem that can be addressed by multi-criteria decision aiding (MCDA) techniques, but most people do not conduct a multi-criteria analysis when facing that choice. The study aims at finding out to what extent consumer preferences can be approximated by a multi-attribute additive model (additive value function). Surveys on stated preferences were conducted by interviewers/analysts trained in this type of models. An initial survey (convenience sample, n=376) allowed defining which criteria should be considered and provided lessons for the construction of a second survey. In this second survey (n=252), subjects were also asked to answer a conjoint analysis type of questionnaire before and after performing MCDA. This presentation reports on some initial conclusions derived from these experiments. Keywords: Marketing, Multi-Attribute Value Theory, Preference Learning. 234 TU4D (contributed) Multi-project scheduling as a MCDM problem Krzeszowska, Bogumila (*) University of Economics in Katowice Trzaskalik, Tadeusz University of Economics in Katowice One of the most important phase in project portfolio management is planning. During this phase tasks are identified and scheduled. Schedule brings information how tasks should be planned over time, during the realization phase of projects. In real-life applications the schedule should optimize not only project finish time but also resource usage and cash flows. It shows that project scheduling is a MCDM problem. Anyway, there are only few papers that consider the project scheduling in such a way. We will consider multi-project scheduling. The purpose of this paper is to build mathematical model of the project portfolio scheduling problem and present it as a MCDM problem. A discussion about decision variables, required constraints and optimization criteria will be provided. Three criteria multi-project scheduling problem will be considered: sum of penalties for projects delay minimization, resource over-usage minimization and NPV maximization. Keywords: Project Scheduling, Multi-Project Scheduling, Multi-Criteria Decision Making. 235 MO3C (contributed) Multi-Attribute Online Reverse Auctions Wallenius, Hannele (*) Aalto University Long, Pham National Economics University Teich, Jeffrey New Mexico State University Wallenius, Jyrki Aalto University Online procurement auctions play an important role in e-Commerce. The basic principles of such auctions originate from the concept of dynamic pricing and a bidding process to set prices. In terms of B-to-B e-Commerce, online reverse auctions have been used by many large corporations. However, concerns relating to online reverse auction usage have been pointed out. One of the major concerns is that online reverse auctions only concentrate on the interests of the buyer, while ignoring the interests of the suppliers. It is true that long-term relationships between buyer and supplier can be damaged, if price is the only priority of the buyer. In order to overcome the criticism, new auction mechanisms that take non-price attributes, such as quality, delivery and payment terms into consideration have been suggested. Several commercial software vendors provide platforms for conducting such auctions known as Multi-Attribute Online Reverse Auctions (MAORAs). Increasing attention is being paid to MAORAs because of their ability to enable buyers to thoroughly compare the relative strength of potential suppliers based on multiple attributes, including but not limited to price. Theoretically, MAORAs are expected to lead to competitive bidding and allocative efficiency. However, contrary to a common belief, online reverse auctions can generate benefits not only for buyers but also suppliers. Via online reverse auctions, suppliers can gain market information, create new markets for better excess capacity management, and attract new customers. The objective of this paper is to gain a deeper understanding of MAORAs, how to implement them and their outcomes. This is achieved through an extensive state-of-the-art literature review of articles published in academic journals. Specific issues related to MAORAs, such as rules, choice of attributes, reservation prices, the role of bid decrement, scoring functions, winner determination, and information architecture are discussed. Keywords: Auctions/Bidding, Multi-Attribute, Reverse Auctions. 236 FR3F (contributed) Multi-Criteria Decision Analysis for Water Supply Infrastructure Planning under Uncertainty Scholten, Lisa (*) Eawag Lienert, Judit Eawag With asset lifetimes ranging from a few decades to more than a century, water supply infrastructure choices are to be made on the long-run. To support long-term planning of water supply systems, multi-attribute utility theory (MAUT) under four future scenarios is applied to a case study in Switzerland. The problem-framing was done in workshops and individual interviews in which stakeholders from different professional backgrounds representing the local, cantonal, and national levels were involved. Altogether, eleven decision alternatives -based on organizational, managerial and technical characteristics- are evaluated under five fundamental objectives, which in turn consist of 30 lower level fundamental objectives. Different sources of uncertainty are often ignored in practical MCDA due to the complexity of elicitation and preference modeling. Several studies addressing these uncertainties exist. Nonetheless, the relative importance of single uncertainty sources for overall uncertainty, and thus the ability to discriminate between alternatives, is usually not investigated. Instead of using the common simplifications (equal weights, linear value functions, risk neutrality, additive aggregation) and ignoring potentially important uncertainties, we 1) elicit weights, value functions, and risk attitude in interval form from ten stakeholders using a simplified procedure, 2) quantify the alternative’s outcomes with regard to the attributes including uncertainty using model predictions and expert judgment, and 3) use this information for a global sensitivity analysis to find out which of the individual uncertainty sources contribute most to overall uncertainty. This will help to focus on more adapted elicitation design and methodic improvements which better account for, and reduce, uncertainty in complex decision support situations. Keywords: Decision Analysis, Multi-Attribute Utility Theory, Multi-Criteria Decision Analysis, Sensitivity Analysis, Uncertainty Modelling, Water Infrastructure Planning, Water Resources Management. 237 MO3B (contributed) Multi-Criteria Decision Making Applications in Higher Open and Distance Learning Systems Kamisli Ozturk, Zehra (*) Anadolu University In this study, we consider a number of the general multi criteria decision problems that may be encounter in many Higher Open and Distance Learning (ODL) institutions, and we also discuss possible solution approaches like mathematical programming approaches, AHP, ANP and genetic algorithms. The terms open and distance learning represent approaches that focus on opening access to education and training provision, freeing learners from the constraints of time and place, and offering flexible learning opportunities to individuals and groups of learners. The management of ODL institutions often differs from the management of traditional face-to-face educational institutions, this being reflected in the structure of the institution and generating a clear differentiation between the types of learners. The management of ODL institutions has a particular character reflecting the wide area served by the education provision. Some of the multi criteria problems faced in ODL systems are vehicle routing, learner-practice assignment, application model selection and learning management system selection. As seen from the examples, managers have to make decision around various tasks continuously. Usually, the effort is towards minimising the operating costs of ODL systems and as technological progress continues these multi objective decision problems will continue to occur. The presence of multiple actors, e.g. academic personnel and learners etc., brings out user satisfaction problems and issues that complicate these decision problems and, furthermore, in educational institutions most of the problems are institution-specific because of particular local constraints and parameters. Under these circumstances, in the solution processes of most problems a need for heuristics occurs. We have looked at a small subset of ODL related decision problems in this study, and many of the problems discussed here can be found across the spectrum of traditional higher education systems. Keywords: Analytic Hierarchy Process, Analytic Network Process, Assignment, Genetic Algorithms, Open And Distance Learning. 238 TU3B (contributed) Multi-Criteria Decision Making Approach for Resource Efficient Bread Supply Chains Banasik, Aleksander (*) Wageningen University Claassen, G.D.H. (Frits) Wageningen University Kanellopoulos, Argyris Wageningen University van der Vorst, Jack Wageningen University Resources are becoming scarce, ecosystems are threatened and people become increasingly aware of the environmental pressure caused by food production. With this changing socio-economic environment, the bread industry in the Netherlands needs to adapt to remain competitive too. Therefore all links of the supply chain pursue for solutions that result in sustainable production and efficient use of resources. The current structure of the Dutch bread industry in relation with the nature of the product (i.e. short self-life) results in large waste streams. It is estimated that in the Netherlands, one third of the total bread production is lost as waste. This research focuses on improving sustainability performance, resource efficiency and effectiveness of the whole chain such that raw materials and resources are used to their full potential. To achieve this, existing inefficiencies should be identified, quantified and eliminated, and opportunities for valorization must be explored. Looking to the chain as a whole enables a more holistic analysis but also increases complexity because of the conflicting objectives in different links of the chain. Moreover, the multiple and to a large degree conflicting dimensions of sustainability (i.e. economic, environmental and social) add to the complexity involved in quantifying and analyzing performance of the bread production chain. In this study, exergy, which is proposed in engineering as a comprehensive indicator of the environmental dimension of sustainability, and advanced quantitative decision support tools is used to optimize production and re-design the logistical structure of the whole chain. A Multi Criteria Decision Making model is proposed to support managerial decisions and improve sustainability performance of the bread food supply chain in the Netherlands. Pareto-optimal solutions and trade-offs between a number of economic and environmental indicators are calculated. Keywords: Exergy, Food Supply Chains, Sustainability. 239 MO4F (invited) Multi-Criteria Decision Making: Exergy and Energy Life Cycle Analyses for Plastic Individual Disposable Bags versus other Material Escobar-Toledo, Carlos Enrique (*) National University of Mexico Alegr´ıa Mej´ıa, Lol-chen University of Mexico Ram´ırez, B´arbara National University of Mexico This paper explores the application of a multi-criteria decision making model within energy utilization in sustainable development and the potential sources to increase energy efficiency in the manufacture of plastic materials used for disposable bags as well as other material used for the substitution of plastic bags. The Exergy loss during the whole life cycle, which lead to the analysis of production routes with better thermodynamic performance and hence less damage to the environment, is presented as a tool in aid of the trade-off of alternative materials in the multi-criteria decision making. Because the comparison between materials and / or production processes is more complete when you look at more than one criteria, are also analyzed other factors besides the irreversibility as: the amount of fossil fuels consumed (expressed as barrels of oil equivalent), the amount of energy consumption (also expressed in barrels of oil equivalent), the amount of emissions of greenhouse gases and the product value as an economic factor. The material analyzed are: polyethylenes of high and low density (HDPE and LDPE), the same with an additive to induce degradation, Kraft-type paper, cotton and polypropylene (PP). Keywords: Decision Making, Energy System, Exergy, Life Cycle Analysis, Multi-Criteria Decision Analysis. 240 WE1D (invited) Multi-Criteria Decision Support for Designing Critical Infrastructure Stress Tests Comes, Tina (*) University of Agder Bertsch, Valentin Karlsruhe Institute of Technology Critical Infrastructures (CIs), such as the energy system, are usually defined by their importance for societies and economy, or by the consequences of their disruptions. Still, there is no unanimous definition of what actually are the most severe consequence that must be avoided. Recent events such as Hurricane Sandy illustrated the lack of preparedness and the need for more robust CIs, particularly with respect to the energy system. As modern societies are characterised by complexity and interlacedness, the impact of failures of CIs cannot be contained regionally or among a small group of stakeholders. Therefore, the need arises for participatory approaches ensuring transparency and facilitating consensus-building. Stress tests for financial or engineered systems have been created to protect population and economy against the consequences of the failure of highconsequence systems. The design of these stress tests typically starts with constructing events (stress test scenarios) that are a means to disturb the system. Usually, stress test scenarios do not cover emerging and unprecedented risk, and there is no systematic way to steer and manage the scenario construction such that it reveals the most critical elements of the system. Criticality is understood in terms of vulnerability (susceptibility to failure) and consequence or harm. We will present an MCDA approach to design stress tests for CIs. For its outstanding importance, we will focus on the energy. This approach focuses on requirements for stress tests, such as protection levels desired or minimum plausibility of stress test scenarios, but also takes into account longer-term goals of economic growth and well-being. In this way, decision-makers are supported in balancing efficiency and effectiveness, i.e., in making trade-offs between optimisation for standard operations and foreseen changes vs. ensuring that service levels can be achieved even when the system is stressed by shocks or unexpected trends. Keywords: Energy System, Participatory Evaluation, Risk Management, Scenarios, Vulnerability. 241 TU3D (contributed) Multi-Criteria Decision Support in Paper Production Industry Claassen, G.D.H. (Frits) (*) Wageningen University Gerdessen, Johanna Wageningen University Pulp and paper industry is an extremely large business characterised huge capital costs. As margins are low, a continuous search for efficient production and decision support is of utmost importance in this branch of industry. Faced with rising raw material and energy costs, pulp and paper producers have to optimise the performance of their processes to remain competitive. Reducing costs, e.g. for raw materials, is an important way to increase or (at least) maintain the annual operating profit. We present the development and continuous improvements of an OR-based Decision Support System (DSS) in paper production industry. The first pilot system was handed over to the company in 1990. After two decades the basic concepts of the initial system are still used on a regular basis. However, pressed by changed circumstances, i.e. the acquisition by a large globally operating company with significantly more mills and the constantly advancing physical knowledge of large scale pulp- and paper production, management decided to upgrade the system to a flexible tool for global decision support on different decision levels within the company. Several features are added, updated and improved. One of the main concerns from practice was that a single overriding objective (i.e. minimizing costs) is too restrictive for future decision support. Nowadays, management assistants focus on a variety of objectives. The aim for a DSS in an MCDM context did have a substantial impact on the latest upgrade in which mixed integer (0-1) fractional programming was introduced. The aim of this contribution is 1) to present the main characteristics of the DSS, 2) to show why an MCDM approach induced a (0-1) fractional programming problem and 3) to demonstrate how various, partly non-linear, and conflicting objectives can be solved such that the added value of the system is recognized and adopted by management assistants in practice. Keywords: Applications of MCDM, Decision Support Systems, Fractional Programming, Mixed Integer Programming, Multi-Criteria Decision Making. 242 MO4G (contributed) Multi-Criteria Decision-Making Perspective on the Privacy versus Security Conundrum Toubaline, Sonia (*) University College London Borrion, Herv´e University College London Le Sage, Tanya University College London Mitchener-Nissen, Timothy University College London In the last decade, many governments have responded to evolving criminal threats by funding and implementing new security policies and technologies. Whilst many of these constitute enhanced versions of existing measures, a number of them were specially designed to accompany emerging technologies and applications such as social networking. As the threat of terrorism and cybercrime continue to rise in our democratic societies, a major challenge for those in charge of developing effective security measures lies in minimizing the ethical issues associated with their implementation. Faced with the possibility of moving towards the ‘surveillance society’, experts and non-experts recurrently take part in debates organised to assess the need for new security measures. However, typical presentation of the problem which often consists of opposing privacy and security is sometimes criticized for inherently supporting securitization of our society. In this article, we analyse the privacy-security conundrum using a multi-criteria decision making framework. We discuss some of the limitations arising from framing the problem as a privacy-security tension, and provide several directions that may be used to better guide and inform the public and policy-makers on matters related to security. The paper concludes by presenting some initial results obtained from an attempt to elicit a clear list of criteria for the specific purpose of assessing security measures. Keywords: Applications of MCDM, Criteria, Decision Making, Privacy, Risk Management. 243 TU5E (contributed) Multi-Criteria IMRT Optimisation based on the Generalised EUD Model Cabrera G., Guillermo (*) The University of Auckland Ehrgott, Matthias Lancaster University Mason, Andrew The University of Auckland Intensity modulated radiation therapy is one of the major forms of cancer treatment. Its main goal is to eradicate all clonogenic cells by delivering ionizing radiation from an external source to the tumour without compromising surrounding normal tissues. To design a treatment plan firstly the number and directions of radiation beams are determined. Then, the intensity of the radiation for each beam angle needs to be computed. Finally, radiation is modulated using a physical device called a multileaf collimator which requires the solution of a sequencing problem. In this talk we address the second problem, known as the fluence map optimisation (FMO) problem. FMO attempts to find a combination of beam intensities that produce a clinically acceptable treatment plan. Several models have been proposed to solve the FMO problem. A well-known measure of the biological effects of the delivered radiation dose is the generalised equivalent uniform dose (gEUD). One advantage of gEUD-based optimisation is that it penalises hot and cold spots in organs at risk (OAR) and target regions, respectively. Therefore, gEUD allows us to evaluate and compare different intensity maps generated during the solution of the FMO problem even though these intensity maps produce non-uniform dose distributions. Based on the fact that gEUD is a positively homogeneous function, we propose a new strategy to solve a multi-objective FMO model. This model attempts to maximise the target (tumour) gEUD while minimising the gEUD of the OARs. Numerical experiments applied to a prostate cancer case with three structures, namely, prostate, bladder and rectum have been carried out. Results show that, based on our proposal, the whole efficient set can be found by following a simple procedure which includes the solution of a convex single-objective optimisation problem which can be solved to optimality using well-known gradient based algorithms. Keywords: Intensity Modulated Radiation Therapy, Multi-Objective Optimization, Positively Homogeneous Functions. 244 TH3C (contributed) Multi-Criteria Preference Disaggregation Analysis for Classification: an Application to Model Colour Preferences Ghaderi, Mohammad (*) Agell, N´ uria ESADE, Universitat Ramon Llull Ruiz, Francisco Javier Sanchez, Monica Universitat Polit`ecnica de Catalunya Classification or discrimination problems consider the assignment of a set of alternatives into predefined groups. In some situations, groups are defined in an ordinal way from the most to the least preferred. In the multi-criteria decision-aid (MCDA) literature, this is known as a ‘sorting’ or ‘learning preferences’ problem. Capturing the decision makers (DMs) tacit knowledge, by providing them a training sample to be sorted in an ordinal way, is considered of interest in the knowledge management field. Extracting and mathematically framing the preference system of the decision maker (expert) enables us to predict preferences for cases that are outside of the training sample. Much effort has been made in this direction in the area of artificial intelligence, specifically in fuzzy set theory and machine learning systems. Preference disaggregation, as one of the most popular approaches for capturing the preference system of DMs, in MCDA is used to infer global preference models from given preferential patterns. Among others, we can highlight: UTA (UTilites Additives); UTASTAR; UTADIS (UTilites Additives Discriminates) ; ELECTRE TRI ; and MHDIS methods. The aim of these approaches is to provide a model that is as consistent as possible with the decisions made by the DM. This research includes a literature review of the existing methodologies for learning preferences and a comparison between some of them. An application related to colour preferences is used to compare these methodologies. Finally, managerial applications involving learning colour preferences are studied. Keywords: Aggregation Schemes, Classification, Decision Maker’s Preferences, Fuzzy Logic, Multi-Criteria Decision Analysis, Preference Disaggregation, Preference Learning. 245 FR3E (contributed) Multi-Criteria Sorting with Category Size Restrictions ¨ Ozpeynirci, Selin (*) ˙ Izmir University of Economics Koksalan, Murat Middle East Technical University Mousseau, Vincent Ecole Centrale Paris In this study, we consider the multi-criteria sorting problem where alternatives that are evaluated on multiple criteria are assigned into ordered categories. Sorting problems arise in many real life situations such as classifying countries into different risk levels based on economical and sociopolitical criteria, evaluating credit applications of bank customers, evaluating student applications for graduate programs and selecting projects for different kinds of funding policies. There are several approaches to the sorting problem in the literature based on estimating a utility function, ordinal classification, outranking relations, and interactive approaches. We focus on the sorting problem with category size restrictions, where the decision maker may have some concerns or constraints on the number of alternatives that should be assigned to some of the categories. We develop an approach based on the UTADIS method that fits an additive utility function to represent the decision maker’s preferences. We introduce additional variables and constraints to enforce the restrictions on the sizes of categories. We demonstrate our approach on two applications from practice. Keywords: Category Size, Multi-Criteria Sorting, UTADIS. 246 WE2F (invited) Multi-Criteria Timetable Information - Extending the Pareto Approach Schnee, Mathias (*) Technische Universitat Darmstadt Keyhani, Mohammad Hossein Technische Universitat Darmstadt Timetable information systems calculate train connections and customers choose among these according to personal preferences. This is a multi-criteria decision in a natural way with profiles varying drastically between customers, e.g. compare a business traveler primarily concerned with travel time to an elderly person focusing on the number of interchanges and ticket cost. To satisfy all types of customers, we developed a timetable information system (MOTIS) based on multi-criteria optimization. An engineered generalization of a Pareto version of Dijkstra’s algorithm on a suitably constructed timeexpanded timetable graph is used to determine attractive connections for all users. The developed models and algorithms are fully realistic and are applied on real data from German Railways. Besides the classical criteria travel time, number of train changes, and ticket cost, we also regard the difference in departure and arrival times, uninterrupted sleeping time in night trains, special offers based on contingents for trains, and the reliability of train changes. This reliability is based on probability distributions for the delays of arrival and departure events. In order to determine all attractive connections, we extended the classical Pareto-approach to a) find additional connections (e.g. only slightly worse but largely differing departure and arrival times) and b) disregard undesired Pareto-optima (e.g. much longer travel time but only slightly less expensive). The system is efficient (less than 2 seconds on average for real customer queries) and finds a set of high quality connections with reasonable size (4-8 optima on average). In this paper, we will formally introduce our extended Pareto-approach and discuss the treatment of the different criteria. We will evaluate the repercussions of the chosen optimization approach (classical Pareto or our different extensions) and will present a study on the optima and run time depending on the selection of criteria. Keywords: Multi-Objective Optimization, Shortest Path, Timetable Information, Train Networks. 247 MO4E (contributed) Multi-Objective Dynamic Scheduling Optimization Model for Real-Time Rebalancing Control of the Hybrid Assembly Lines Wong, Wai Keung (*) Hong Kong Polytechnic University This paper investigates the dynamic scheduling optimization problem (DSOP) of hybrid assembly lines in apparel manufacturing. A reliable, flexible and effective dynamic scheduling system was proposed for rebalancing control of real-time hybrid assembly lines. There are three stages in our proposed approach. First, the updated radio frequency identification-based (RFID) data capture system is adopted to collect real-time production records, which provides a guarantee for rebalancing control to react to disturbances of production conditions accurately and timely. Second, a discrete event simulation model is developed to serve as dynamics of hybrid assembly lines, which solves the difficulty of calculating objective functions when no closed-form mathematical formulation exists. The third stage is to establish a dynamic scheduling optimization model for rebalancing control of hybrid assembly lines. An enhanced discrete differential evolution (EDDE) algorithm was proposed to find an optimal solution to the DSOP. The EDDE algorithm was enhanced by integrating heuristic selection strategies to yield a satisfactory solution as quickly as possible and repairing strategies to handle constraints effectively. The results from the numerical experiments verified the feasibility and effectiveness of the proposed EDDE algorithm in finding optimal solutions to the DSOP. Compared with industrial practice, our proposed system can improve production efficiency with an increment of above 17%, which leads to reduction in production time of about 15% for orders of the same size, which demonstrates that the proposed dynamic scheduling system is a promising approach to improvement in efficiency and productivity of hybrid assembly lines by effective rebalancing control. After successful validation, the proposed model is now practically used by the manufacturers in the apparel industry in Mainland China. Keywords: Line Balancing Control, Multi-Objective Optimization, Operator Allocation, Production. 248 TH3E (invited) Multi-Objective Framework for Model Based Design of Experiments: Application to Hemodialysis and Type 1 Diabetes Model Maheshwari, Vaibhav (*) Rangaiah, Gade Pandu Samavedham, Lakshminarayanan National University of Singapore Model-based design of experiments (MBDoE) has grown into an indispensable tool for improving parameter precision of dynamic models. In the last two decades, the applications of MBDoE have extended from chemical kinetics to biomedical applications, biological networks, biological processes (fermentation and bio-kinetics), environmental engineering, etc. Traditionally, alphabetical design D-, A-, and E-design criteria are used for model-based experimental design, where each design maximizes the information measure for improving precision of parameters. The information measure is calculated from Fisher information matrix (FIM). However, it is observed that maximizing the information measure, inevitably leads to increase in correlation among parameters. This increased correlation is associated with poor point estimates and/or reduced precision of parameters. To overcome the weakness of traditional alphabetical designs, a multi-objective optimization (MOO) based framework for experimental design has been proposed. The proposed framework consists of two contradicting objectives, namely, (1) maximize the information measure (as in traditional alphabetical design), and (2) minimize the correlation among parameters. In this work, we illustrate the application of proposed experimental design framework for a toxin kinetic model in hemodialysis and Type 1 diabetes model, and compare its performance with alphabetical designs. For each of D-, A-, and E-designs, corresponding DMOO, AMOO, and EMOO is proposed and its performance is assessed for parameter precision and correlation. Obtained results indicate that MOO based experimental design framework provides improved point-estimates when compared with those obtained from alphabetical design. Acknowledgement: We thank National Kidney Foundation, Singapore for supporting this research through NKFRC/2011/01/06. Keywords: Correlation, Model-Based Design Of Experiments, Multi-Objective Optimization, Parameter Precision, Point Estimates. 249 MO4B (invited) Multi-Objective Modelling and Simulation in New Product Development (NPD) using the Belief Rule-Based (BRB) Methodology Savan, Emanuel (*) The University of Manchester Chen, Yu-Wang The University of Manchester Xu, Ling The University of Manchester Yang, Jian-Bo The University of Manchester In this paper, the novel belief rule-based (BRB) methodology is employed to identify the underlying relationships between instrumental measurements and their corresponding sensory attributes. The developed BRB model is then used to support the decision making (DM) process in new product development. As a case study, data collected from Unilever on consumer perception of 70 products over 55 sensory attributes has been used along with over 100 instrumental measurements. In the modelling phase, one of the most prominent interests raised by the practitioners concerned the model’s validation and overfitting. Given the recent development of the BRB methodology, this model building phase has not yet been formally addressed. Consequently, one of the main purposes of this paper is to develop a validation framework capable of indicating the appropriate model structure. In order to guarantee the comprehensiveness of the carried validation, the proposed framework was built based on three complementary criteria. The Akaike information criterion is used in conjunction with another well recognised approach: the Bayesian information criterion, and with a heuristic approach, in order to identify the most suitable model structure. Various possible structures are generated following a compilation of search heuristics: genetic algorithms, tabu search and a heuristic expert-knowledge based search. The purpose of this research is twofold. On one hand, the methodology is applied to a new, complex, DM context. It is compared with Unilever’s initial model, as well as with the most popular methodologies in the field: artificial neural networks, support vector machines, and multiple linear regression. It is interesting to note that the BRB methodology outperforms the other approaches in the case study, proving highly suitable for the addressed DM context. On the other hand, a validation framework is developed and analysed. The proposed model structure proved highly robust upon testing. Keywords: Belief Rule-Based methodology, Model Testing And Validation, New Product Development, Over-Fitting, Sensory Attributes Prediction. 250 TU2G (invited) Multi-Objective Optimisation for the Modeling and Control of Uncertain Dynamic (Bio)Chemical Processes Logist, Filip (*) Coenen, Thomas Stamati, Ioanna Telen, Dries Vallerio, Mattia Van Impe, Jan Katholieke Universiteit Leuven Many (bio)chemical processes are dynamic in nature and their optimisation involves the determination of optimal time-varying control policies. Moreover, in view of sustainable development multiple objectives arise. For instance not only more profitable (economically sustainable), but also cleaner (environmentally sustainable) and safer processes (socially sustainable) processes are aimed at. To enhance decision making in practice, it is important to enable the efficient generation of optimal alternatives (Pareto optimal solutions) and the evaluation of the trade-offs between them. Recently a freely available toolkit ACADO (www.acadotoolkit.org) for multi-objective optimal control of dynamic systems has been introduced. The rationale behind the approaches is a tight integration of efficient multi-objective scalarisation strategies (e.g., Normal Boundary Intersection and Normalised Normal Constraint) with fast deterministic dynamic optimisation approaches (e.g., Single and Multiple Shooting). Some recent developments will be illustrated for different (bio)chemical applications. First, results are presented for the multi-objective optimisation under uncertainty of fed-batch fermentors for the production of lysine. Second, multi-objective optimisation is applied to the experimental design of dynamic experiments for modeling the temperature dependency of the growth rate of micro-organisms. Third, a case study is discussed regarding the tuning of nonlinear model predictive controllers in view of grade changes in a large-scale chemical plant. For each of the three cases the Pareto set as well as the corresponding optimal operation policies are presented. Keywords: Applications of MCDM, Chemmical Engineering, Control, Dynamic Optimization, Dynamic Systems, Multi-Objective Optimization, Optimal Control. 251 FR3D (contributed) Multi-Objective Optimization of Combinatorial Problems with Fuzzy Data Bahri, Oumayma (*) ISG Tunis - University of Tunis Ben Amor, Nahla ISG Tunis El-Ghazali, Talbi University of Lille Optimization under uncertainty is an important line of research, having today many successful real applications in different areas. Despite its importance, few works on multiobjective optimization under uncertainty exists today. However, the panoply of existing studies has been often experimented on deterministic multi-objective problems or even on uncertain multi-objective case but by treating it as mono-objective (in considering the set of objectives as if there is only one). The aim of our study is to deal with such a multi-objective problem with uncertain data, using the possibility theory which offers a flexible and efficient approach for handling uncertainty. So, we propose firstly a new adaptation of the possibilistic framework, in which the uncertain data are represented by a specific form of possibility distributions, commonly known as triangular fuzzy numbers. Then, knowing that the main goal of solving a multi-objective problem is to identify the set of optimal solutions called ’Pareto optimal set’, the result obtained in our case will not be a set of exact solutions but a set of uncertain solutions disrupted by the used fuzzy form. Clearly, the obtained fuzzy solutions cannot be classified by using the classical Pareto dominance. Thus, we propose at the second stage a new Pareto dominance for ranking triangular fuzzy numbers. The proposed method is then applied to solve a multiobjective variant of Vehicle Routing Problem (VRP) with uncertain customer demands. Finally, to validate our study, two well known multi-objective algorithms: PLS1 (Pareto Local Search Algorithm) and SEEA (Simple Elitist Evolutionary Algorithm) are adopted and then developed under the software platform dedicated to multi-objective optimization: ParadisEO-MOEO. The both algorithms are tested on a few instances of Solomon’s benchmark and evaluated by using two quality metrics: Hypervolume and Epsilon indicators. Keywords: Multi-Objective Optimization, ParadisEO-MOEO, Pareto Dominance , Pareto Local Search Algorithm, Possibilistic Framework, Possibility Theory, Simple Elitist Evolutionary Algorithm, Triangular Fuzzy Numbers, Uncertainty, Vehicle Routing Problem. 252 TU2G (invited) Multi-Objective Optimization of Integrated Design and Control of a Paper Mill Steponavice, Ingrida (*) University of Jyv¨askyl¨a Miettinen, Kaisa University of Jyvaskyla Many real-life engineering design problems involve an extensive use of dynamic process simulations. Engineers rely on these simulations to develop new products, better understand experimental results and build confidence in solutions they develop. Solving dynamic problems is very demanding due to typically high computational cost, uncertainty and intractability related to dynamic process simulation. Moreover, it is often insufficient to only consider a single objective function to be optimized. Instead, many optimal design problems have multiple conflicting objective functions that must be optimized simultaneously. Multiobjective optimization problems have multiple Pareto optimal solutions which are mathematically equal and a human decision maker is necessary to determine the most desirable one. With an increasing complexity of design problems, finding an optimal design in real-life applications remains a challenging task. We present an efficient approach to search for optimal solutions of such problems enabling an effective decision making process allowing learning and exploring different trade-off solutions. We demonstrate our approach with a case study in paper making industry. To be more specific, we consider both optimal design and control of a paper mill, where we aim at decreasing the investment cost and enhancing the quality of paper on the design level and, at the same time, guaranteeing the smooth performance of the production system on the operational level. Keywords: Applications of MCDM, Computational Cost, Decision Making, Pareto Optimal Solutions. 253 MO5H (poster) Multi-Objective Optimization Problems in Statistical Machine Translation Duh, Kevin (*) Nara Institute of Science and Technology Sankaran, Baskaran Simon Fraser University Sarkar, Anoop Simon Fraser University Machine Translation has been one of the holy grails of Artificial Intelligence, the goal being to convert a sentence in one language to another language while preserving the meaning. Thanks to the increasing availability of large textual datasets and the increasing performance of statistical optimization algorithms, the field has advanced significantly in the past decade. Many commercial systems, such as those from Microsoft and Google, now employ the statistical approach when developing the translation system. Often, these systems are built by single-objective optimization methods. The objective is the quality of the translation, as defined by some type of string matching metric between the system’s translation and a human translator’s sentence. However, translation quality is a slippery concept and many types of matching metrics based on different linguistic insights have been proposed. Thus we argue that the best approach is to optimize the system on these multiple objectives concurrently. In this work, we describe a formulation of the multiobjective optimization problem in Statistical Machine Translation. We will also discuss a simple solution based on incorporating Pareto Frontier ideas into a Support Vector Machine classifier. The purpose of this presentation is to introduce this novel application of multi-objective optimization to the MCDM community; we hope to learn about some new solution techniques in return. Keywords: Applications of MCDM, Artificial Intelligence. 254 MO2F (invited) Multi-Objective Reinforcement Learning using Sets of Pareto Dominating Policies Van Moffaert, Kristof (*) Vrije Univesiteit Brussel Drugan, Madalina M. Vrije Univesiteit Brussel Now´e, Ann Vrije Univesiteit Brussel Many real-world problems involve the optimization of multiple, possible conflicting, objectives. Multi-objective reinforcement learning (MORL) is an extension to standard reinforcement learning where the scalar reward signal is extended to multiple feedback signals, i.e. one for each objective. Thus, MORL is the process of learning policies that optimize multiple criteria at the same time. We aim to combine multi-criteria decision making (MCDM) and evolutionary multiobjective optimization (EMO) techniques to construct hybrid MORL algorithms. In this paper, we present a novel algorithm that integrates the Pareto dominance relationship into a reinforcement learning approach. This algorithm is called Pareto Q-learning and it is an extension to the single-objective Q-learning algorithm that keeps track of sets of Pareto dominating policies. We also propose two mechanisms for using Pareto Q-learning in an on-line setting, i.e. when the cost of sampling actions is crucial and the agent gradually improves its performance. More precisely, these mechanisms use the hypervolume measure and Pareto dominance relationship to select the most promising actions. We test the algorithm on multiple environments with two and three objectives. We demonstrate that Pareto Q-learning outperforms current state-of-the-art MORL algorithms and is able to find the entire Pareto front on two benchmark instances. Keywords: Multi-Criteria Decision Analysis, Multi-Objective Optimization, Pareto Set, Reinforcement Learning. 255 TU2D (invited) Multi-Objective Restricted Dynamic Vehicle Routing Problem with Time Windows De Armas, J´esica (*) Universidad de La Laguna Meli´an, Bel´en Univesidad de La Laguna Moreno P´erez, Jos´e A. Universidad de La Laguna The Dynamic Vehicle Routing Problem with Time Windows (DVRPTW) consists not only in finding a set of optimal routes for some vehicles in order to serve a set of customers with specific time windows, but also in readjusting the vehicle routes to add new customers that appear in real time. We tackle a real-world application of this problem proposed by a delivery company, in which the vehicles are loaded at the beginning of the planning horizon and then the arising services are inserted into a route supposing the minimum infeasibility. The company knows all the deliveries, but only the 60% of the pick-ups. The remaining 40% is dynamic since it appears during the day. Moreover, the following constraints have to be considered. The vehicles have different capacities; each of them can have a different time window, as well as the customers; extra hours for the vehicles might be allowed; the customers have a level of priority so that they can be postponed if they have less level of priority than a fixed one; and finally, some customers cannot be served by certain vehicles due to road restrictions. Besides, the problem is zoned; a problem supposes the solution of different sub-problems. The objective functions have been taken into account in lexicographic order, in such a way that if there is a draw between two different plans for one objective, the next one is compared. The order used in this work is: number of postponed customers, extra time used by the vehicles, total traveled distance, number of routes, total consumption and total cost. The computational experiments have been carried out by testing the approach over the test problem instances in the related literature. Thus, we have obtained competitive results if we compare them with the best known results. Furthermore, we have generated and tested instances specifying the constraints of the real application. Finally, nowadays we are doing real tests using the 4GFlota system, which allows managing a fleet of vehicles. Keywords: Dynamic VRPTW, Multi-Objective Optimization. 256 TH4F (contributed) Multi-Objective Task Allocation using Multi-Agent Coalition Formation Ben Abdelaziz, Fouad (*) Rouen Business School Amer, Noha American University of Sharjah In this paper, we address the multi-objective task allocation problem. We consider a set of agents each with a vector of limited and/or negative resources that shall be cooperating in order to perform a set of tasks. The cooperation leads to a numerous number of coalitions. In addition, We consider two objectives; maximizing the efficiency of the system and maximizing the agent’s payoff. These two objectives represent two different types of agent’s behaviors; cooperative agents and selfish agents. To solve this problem, we design a genetic algorithm and we compare the obtained results with those obtained by an exact algorithm for small sized problems. We also address the hybrid behavior where agents aim to maximize the efficiency of the system and their own payoff at the same time. We also conclude about the influence of agents’ behavior regarding the performance of task allocation. Keywords: Coalition Formation, Genetic Algorithms, Multi-Objective Optimization, Task Allocation. 257 TU4E (contributed) Multiattribute Utility Analysis using Strict Preference Relations Elicited from a Decision Maker Ichiro, Nishizaki (*) Hiroshima University Masakazu, Ohmi Hiroshima University Tomohiro, Hayashida Hiroshima University To identify multiattribute utility functions with the mutual utility independence, decision makers must specify indifference points and subjective probabilities precisely. By relaxing such exact evaluation, multiattribute value models with incomplete information have been developed. In this paper, we present a new method finding a best alternative through strict preference relations derived from the decision makers by questions that are not difficult to answer. For given strict preference relations, in our method alternatives satisfying the relations can be obtained by solving a mathematical programming problem with constraints representing the preference relations. If there are multiple best alternatives, after eliciting additional strict preference relations from the decision makers, we can eventually find a unique best one by solving the mathematical programming problem with additional preference constraints. Keywords: Multi-Criteria Decision Making. 258 TH4D (contributed) Multicriteria 0-1 Knapsack Problems with K-min Objectives Rong, Aiying (*) Technical University of Lisbon Figueira, Jose Technical University of Lisbon Klamroth, Kathrin University of Wuppertal This paper studies the multicriteria 0-1 knapsack problem (KP) with k-min objectives (MkMIN-KP) in which the first objective is of classical sum type and the remaining objectives are k-min objective functions. The k-min objectives are ordinal objectives, aiming at the maximization of the kth smallest objective coefficient in any feasible knapsack solution with at least k items in the knapsack. The k-min objectives can be treated as a generalized bottleneck objective. From a practical point of view, the MkMIN-KP is an extension of KP by explicitly considering the ranking measures as objective criteria. The ranking measures can be used to model the preferences of the decision makers or the objectives that are difficult to handle by traditional means. Next, we develop efficient hybrid algorithms to obtain the complete non-dominated set of MkMIN-KP based on a variant of the epsilon-constraint method by solving a series of single objective 0-1 multidimensional KPs with a special structure. Keywords: Hybrid Approaches, K-Min Objectives, Multi-Dimensional Knapsack Problem, Multi-Objective Optimization. 259 MO3C (contributed) Multicriteria Cognitive Map: an Integrated Tool for Building a Multicriteria Evaluation Model Cipriano Rodrigues, Teresa (*) Technical University of Lisbon Oliveira, M´onica Duarte University of Lisbon Multicriteria value measurement tools have been extensively used to build evaluation models, which often take a given set of alternatives and evaluation criteria as the starting-point in model building. Nevertheless, it has been recognized that a very demanding and challenging task when building an additive evaluation model is to help the decision-maker to structure the model, in particular constructing a set of criteria and associated descriptors to be used in the evaluation of a set of decision alternatives. Few tools address, in a coherent and integrated way, problem structuring and multicriteria evaluation of decision alternatives. This study proposes a tool – a Multicriteria Cognitive Map (MCM) – which, by ameliorating the capability of making inferences in a cognitive map, supports the construction of a multicriteria evaluation model. Developing a MCM integrates two phases: a first phase to structure the problem by capturing the issues and their systemic relationships in a means-ends network; and a second phase to determine the impact that alternatives (means) have in the values (ends) of the decision-makers. This latter phase uses the MACBETH (Measuring Attractiveness by a Categorical Based Evaluation Technique) approach to measure the strength of the influence of each means-end link. A case study based on the construction of a health index is used to illustrate how the MCM tool can assist building a multicriteria evaluation model. Keywords: Cognitive Map, MACBETH, Multi-Criteria Evaluation Model, Problem Structuring. 260 TU5G (invited) Multicriteria Evacuation Plan for Natural Disasters Ndiaye, Ismaila Abderhamane (*) Tours University Neron, Emmanuel Tours University Since the 90s, the problems of evacuation of persons have been extensively studied in the literature. The proposed models are classified intotwo main categories: macroscopic and microscopic models. The DSS EVAC LOGISTIQUE project is interested in the evacuation of people in the context of flooding, burning or seismic event for which insecurity, capacity and time of crossing roads vary over time. Considering as solved the shelter locations problem, we develop discrete macroscopic models and methods incorporating a notion of risk inherent of the context studied for the evacuation of persons. The problem will be to determine a minimum overall evacuation time while minimizing the risk incurred by each evacuee. In this context, we propose two methods based on Shortest augmenting Paths without using a time-expended network. A efficient exact pseudopolynomial algorithm to solve the Earliest Arrival Flow and the Quickest Flows problems with time dependent data in graphs with one strongly connected component. An epsilonconstrains method for solving Quickest/Safest flows problem with time-dependent data. This research has been supported by the French National Agency of Research ANR-11SECU-002-01- project DSS EVAC LOGISTIQUE. Keywords: Dynamic Network, Earliest Arrival Flow, Evacuation, Macroscopic Model, Multi-Objective Optimization, Quickest Flow, Safest Flow, Time Dependent Data, Transshipment. 261 FR3B (contributed) Multicriteria Model for Locating Distribution Centers Manotas, Luis A. (*) Business Vice-President Manotas, Diego F. Universidad del Valle Ocampo, Germ´an Universidad del Valle Vidal, Carlos Universidad del Valle The problem of location of distribution centers is one of the most important strategic decisions in the field of supply chain management. Usually, this decision has been analyzed with qualitative tools like opinion of experts. In this paper, we propose an integrated framework between real options analysis and multi-criteria decision methods. The idea is to include the strategic value based on expansion options implied in this kind of decisions like a one of the decision criteria to consider. The paper is supported on three subjects: Models for locating of distribution centers, financial models applied to facilities location problems and multi-criteria decision techniques. The methodology developed in this paper includes six types of criteria: financial risk (probability of success) , strategic value (Expansion option), proximity to customers, proximity to production facilities, infrastructure and exposition to natural disasters. The model considered three different location alternatives. To calculate the probability of success and the volatility parameter of each possible location , we used Monte Carlo simulation. After that, we applied the binomial model to value the possible expansion options implied in each location alternative. The multi-criteria method utilized in this work was an AHP-TOPSIS hybrid model. AHP was used to obtain the weights of each criterion and TOPSIS was applied to build a ranking of all location alternatives. The main results show, if the options of expansion are not considered, the best alternative from an economic point of view would be the alternative without expansion option, by his better probability of success, but considering the real options associated with expansion the results changed and the best choices were the alternatives with expansion possibilities. Keywords: Distribution Centers Location Problem, Real Options Analysis, TOPSIS. 262 FR3B (contributed) Multicriteria Model for Selecting Renewable Energy Projects Manotas , Diego F. (*) Universidad del Valle Lozano, Carlos A. Universidad del Valle Vidal, Carlos Universidad del Valle Energy systems play a very relevant role in the social and economic development of a country. In this context, it is necessary to analyze the generation expansion planning (GEP) problem, because decision makers in the energy sector require robust tools to balance the energy demand and the power generation supply. Traditionally, the GEP problem has been analyzed with economic techniques such as levelized cost. However, this is not enough because this kind of problems is influenced by other criteria. One of the most important tasks related with the GEP problem is the selection process among different alternatives for the expansion of renewable energy sources. The problem under study is very complex due to the consideration of different kinds of criteria such as environmental, social, technical and economic factors. The multiplicity of criteria and the involvement of different actors in the decision procedure make multi-criteria analysis a valuable tool in the policy formulation for fossil fuel energy replacement by renewable energy sources. In this paper, we introduce a multi-criteria hybrid model using the traditional Analytical Hierarchy Process (AHP), the Technique for Order Preference by Similarity to Ideal Solutions (TOPSIS), and considering financial risk indicators to prioritize a set of different renewable energy projects in the Colombian electricity market. The methodology used in this work begins with the definition of criteria. At this stage, we review the literature related with the application of multi-criteria analysis to define project portfolios in different types of engineering projects. Later, in order to consider the uncertainty in the decision making process, we include some financial risk indicators such as the Value at Risk (VaR) and the Conditional Value at Risk (CVaR). Finally, we present some results about the use of an integrated framework between financial risk indicators and multicriteria methods. Keywords: Analytic Hierarchy Process, Electricity Production Technologies, Risk Analysis, TOPSIS. 263 WE2E (contributed) Multicriteria Modeling and Optimization of a Market Place of Leads Manel, Maamar (*) Ecole Centrale Paris Mousseau, Vincent Ecole Centrale Paris Ouerdane, Wassila Ecole Central Paris Place des Leads (PdL) is a market place of ’lead’.This one faces a complex problem namely ’the assignment of the leads to the customers’.On the one hand,PdL receives a significant amount of leads in real time.On the other hand,the customers of PdL make orders and can filter the leads that they wish to receive according to multiple criteria.Unlike the classical assignment problems in the literature,the challenge with this problem is to take into account at the same time the optimization of multiple objectives and the real time aspect as well.This work has the aim to tackle this problem.To do that,we propose to formulate this problem as a multi-objective mathematical program.After that,we will consider a stochastic model of the offer and request in the optimization step in order to take into account the real time aspect and therefore the fact that the flows of offer and request evolve continuously.More specifically,a first step is to consider the problem from a single objective and to see afterwards to what extent it is possible to extend it to multiple objectives.Thus,the assignment problem can be described by n orders and m leads,such that the assignment of a lead to a given order is represented by a binary variable xij . Therefore, the set X = {xij /i ∈ {1, . . . , n}, j ∈ {1, . . . , m}} is all possible combinations of leads and orders. However,in our specific problem, we can note that a given order does not accept any lead.In fact,a lead can be assigned to a given order only if it satisfies some criteria called filters. Thus, some pairs xij are not possible.Moreover,some orders require exclusive leads.Thus,we propose an algorithm of pretreatment which constructs two subsets:Xe which represents all xij such as i is an exclusive order and Xe’,the set of xij such as i is a nonexclusive order. In addition,this algorithm checks whether a lead corresponds to a command according to the filters.This will allow us to reduce the research space. Finally,the problem is to find the optimal assignment. Keywords: Assignment, Multi-Objective Optimization. 264 MO4F (invited) Multicriteria Models Applied to Sustainable Enviroment and Efficient Use of Energy Alegr´ıa Mej´ıa, Lol-chen (*) University of Mexico Escobar-Toledo, Carlos Enrique National University of Mexico In a recent research about economic and environmental evaluation scenarios to 2030, based on the potential for GHG in the Mexican energy system, we have included alternate sources of energy for power generation as a case study. We propose in this presentation, a planning approach considering different new technological alternatives from 2025 until 2040, using multiple criteria analysis and mathematical programming to satisfy the demand every five years. Results are analyzed considering also a sensitivity analysis in which is performed a Markov chain analysis. Keywords: Electricity Production Technologies, Markov chains, Mathematical programming, PROMETHEE. 265 MO3B (contributed) Multiobjective Evolutionary Approach to Preference Elicitation in the Analytic Hierarchy Process Mikhailov, Ludmil (*) The University of Manchester The paper is concerned with the problem of deriving priority vectors from pairwise comparison matrices, in the framework of the Analytic Hierarchy Process (AHP). With the exception of the traditional Eigenvector method, most methods for deriving priorities in the AHP are based on some optimization approach. All optimization methods for prioritization introduce a single objective function, which measures the degree of approximation or the distance between the initial judgments and the solution ratios. Then the problem of priority elicitation is formulated as a single-objective optimization task of minimizing the objective function, subject to normalization and some additional constraints, and obtain a single priority vector from a given comparison matrix. A new approach to prioritization in the AHP, which represents the prioritization task as a multiobjective optimization problem was proposed recently by the author. A two-objective prioritisation (TOP) problem is formulated as an optimisation task for minimisation of the Euclidian norm and the number of rank violations, which measure the most important properties of the solutions. The paper investigates some evolutionary and numerical approaches for solving the TOP problem. In order to eliminate the drawbacks of the numerical methods, we propose two evolutionary solution approaches. In the first evolutionary approach, the TOP problem is transformed into a single-objective one, which is then solved by a standard evolutionary algorithm (EA). We also investigate the application of a multiobjective evolutionary algorithm (MEA) for solving the TOP problem. The numerical results show that the MEA outperforms the gradient search method and the single-objective EA with respect to accuracy and computational efficiency. Keywords: Analytic Hierarchy Process, Evolutionary Multi-Objective Optimization, Multi-Objective Optimization, Prioritisation Methods. 266 TU2F (contributed) Multiobjective Land-Use Planning Based on Proximity Measures Stewart, Theodor (*) University of Cape Town We have previously discussed the problem of allocating different land uses across a region, when an important concern is spatial contiguity of activities. These problems lie between the two extreme cases which are simpler to solve than the general problem, namely optimization of land use for each parcel independently of others (no contiguity constraints), and partitioning of the region into single contiguous areas for all land uses. In the intermediate case, the number of contiguous areas for each land use is not predetermined, but the degrees of contiguity of each land use become decision criteria (in addition to the usual criteria related to economic, environmental and social issues. Our original work was based on a grid network for which efficient algorithms could be developed. The number of grid cells needed to model the desired resolution at all points in the region became prohibitively large for rapid solution in interactive systems. The alternative was to use a vector-based representation of the region in terms of polygons of varying sizes. This interfaced well with GIS systems in which the data were stored. This approach was in principle much more efficient, but required users to extract quite complicated information from the GIS, which can also be time consuming and for which requisite skills are not always available. In this paper we discuss an alternative use of vector-based data, but in which the algorithm looks only at distances between polygons. Such data is easily extracted from the GIS. We present numerical results to show that this simpler approach, linked to a specially designed genetic algorithm, produces results virtually indistinguishable from thos based on the original vector-based algorithm. Keywords: Genetic Algorithms, Land Use Planning, Reference Point. 267 TU5E (contributed) Multiobjective Optimization of High-Performance Reinforced Concrete I-Beams by Simulated Annealing Yepes, V´ıctor (*) Universitat Polit`ecnica de Val`encia Alcal´a, Juli´an Universitat Polit`ecnica de Val`encia Garc´ıa-Segura, Tatiana Universitat Polit`ecnica de Val`encia The present study aims to describe a methodology to design reinforced concrete I-Beams based on multiobjective simulated annealing (MOSA) algorithm applied to three objective functions, namely, the economic cost, the durability, and the overall safety of the structure. This methodology was applied to a 15 m beam span for several high-performance concrete mix compositions. The evaluation of solutions follows the Spanish Code for structural concrete. The solution of this simply supported concrete I-beam is defined by a total of 20 variables. Pareto results of the MOSA algorithm suggest the feasibility of more economic structures remaining in service for long periods and ensuring safety. This methodology is widely applicable to a different structure designs and therefore, gives to engineers a worthy guide of the benefits of economic, durable and safety perspectives. Further, the methodology proposed will help structural engineers to enhance their designs of reinforced concrete structures. Keywords: Durability, Metaheuristics, Multi-Objective Optimization, Simulated Annealing, Structural Modeling. 268 TU4D (invited) Multiobjective Optimization with Combined Global and Local Metamodeling Azarm, Shapour (*) University of Maryland Hu, Weiwei University of Maryland Saleh, Khaled University of Maryland Approximation Assisted Optimization (AAO) is widely used in engineering design problems to replace computationally intensive simulations with metamodeling. Traditional AAO approaches employ global metamodeling for exploring an entire design space. Recent research works in AAO report on using local metamodeling to focus on promising regions of the design space. However, very limited works have been reported that combine local and global metamodeling within AAO. In this presentation, a new approximation assisted multiobjective optimization approach is developed. In the approach, both global and local metamodels for objective and constraint functions are used. The approach starts with global metamodels for objective and constraint functions and using them it selects the most promising points from a large number of randomly generated points. These selected points are then “observed”, which means their actual objective/constraint function values are computed. Based on these values, the “best” points are grouped in multiple clustered regions in the design space and then local metamodels of objective/constraint functions are constructed in each region. All observed points are also used to iteratively update the metamodels. In this way, the predictive capabilities of the metamodels are progressively improved as the optimizer approaches the Pareto optimum frontier. An advantage of the proposed approach is that the most promising points are observed and that there is no need to verify the final solutions separately. Several numerical examples are used to compare the proposed approach with previous approaches in the literature. Additionally, the proposed approach is applied to a CFD-based engineering design example. It is found that the proposed approach is able to estimate Pareto optimum points reasonably well while significantly reducing the number of function evaluations. Keywords: Computationally Expensive Problems, Multi-Objective Optimization. 269 TU4D (contributed) Multiobjective Variational Problems Involving Generalized Higher Order Functions Shukla, Kalpana (*) Banaras Hindu University We consider the following multiobjective variational problem as the following form: Z min b Z 1 f (t, x(t), x(t), ˙ x¨(t))dt, . . . , min a b f k (t, x(t), x(t), ˙ x¨(t))dt a subject to x(a) = 0 = x(b), x(a) ˙ = 0 = x(b), ˙ hj (t, x(t), x(t), ˙ x¨(t)) ≤ 0, j ∈ M ≡ {1, 2, . . . , m}, x ∈ P S(T, Rn ), where functions f i , i ∈ K = {1, 2, . . . , k} and hj , j ∈ M = {1, 2, . . . , M } are continuously differentiable functions defined on I × Rn × Rn × Rn . In this paper we have established some optimality conditions for the multiobjective variational programming problems with generalized convexity of higher orders. A higher order dual is associated and weak and strong duality results are established under generalized convexity assumptions. References: [1] Aghezzaf, B., Khazafi, K. (2004): Sufficient optimality conditions and duality in multiobjective variational programming problems with generalized invexity; J. Control Cyber., Vol. 33, pp. 1-14. [2] Bhatia, D. and Kumar, P. (1996): Duality for variational problems with vex functions; Optimization, 36, 347-360. Keywords: Daily Used Materials, Duality, Multi-Objective Optimization, Variational Problems. 270 TH4F (contributed) Multiple Criteria Decision Making in Anti-Collision Scheduling for Level Crossing Tasks Levchenkov, Anatoly (*) Riga Technical University The paper proposes the solutions for applying the MCDM, embedded devices, schedule theory and immune algorithms to minimize simultaneous existence of a railway transport and auto vehicle at a crossing, i.e. avoiding a possibility of collision. Nowadays transport flows are increasing. The traffic congestion and accidents become an everyday reality. It is hard to evaluate losses from the accidents and all these consequences for the state in general and a separate citizen, but the negative effect is definitely proved. More serious consequences are colliding railway transport with auto vehicle. The main goal of the research is to investigate and develop the algorithms of MCDM for road and railway transport scheduling and control. The research includes the analysis of existing railway and city transport system, the development of the mathematical models and multiple criteria schedule affinity assessment function of scheduling immune algorithms, antigen and antibodies models and for anti-collision task and investigation. The computer model of sequence of level-crossings with railway and road transport flows with stochastic parameters, such as traffic, technical condition, weather conditions, is created. Also prototypes of embedded intelligent devices are manufactured to test is in real conditions. Set of more than thousands experiments have been performed. The developed models and algorithms may improve the safety level of transport system control. The schedules developed by the immune algorithm provide better results than the original schedule in 100% cases by all target function criteria. Essential reduction of collision probability is possible from the original value 0.94 to result value 0.01. The research leading to these results has received funding from the ARTEMIS Joint Undertaking and from Latvian Academy of Science and Ministry of Education and Science under grant agreement n◦ 295373. Keywords: Embedded Systems, Multi-Criteria Decision Making, Scheduling. 271 TH3B (invited) Multiple Criteria Hierarchy Process for ELECTRE TRI Methods Corrente, Salvatore (*) University of Catania Greco, Salvatore University of Catania Slowinski, Roman Poznan University of Technology We are considering the hierarchy of criteria in ELECTRE TRI methods designed for multicriteria sorting problems. We discuss several possible options for this generalization, in particular: - with respect to construction of the outranking relation in the context of hierarchical criteria, we shall consider: o outranking relation built in the way of ELECTRE IS method, o outranking relation built in the way of ELECTRE III method; - with respect to reference profiles, we shall consider: o boundary reference profiles separating consecutive classes (ELECTRE Tri-B), o central reference profiles characterizing particular classes (ELECTRE Tri-C); - with respect to the number of reference profiles, we shall consider: o single and multiple boundary reference profiles (ELECTRE Tri-B), o single and multiple central reference profiles (ELECTRE Tri-C, ELECTRE Tri-nC); - with respect to the sorting procedure for ELECTRE Tri-B, we shall consider: o the optimistic procedure, o the pessimistic procedure; - with respect to interaction among criteria, we shall consider: o the classical concordance index without interaction among criteria, o the concordance index with possibility of mutual strengthening effect, mutual weakening effect and antagonistic effect, o effects of reinforced preference and counter-veto in credibility of outranking. We shall also discuss inference of parameters using Robust Ordinal Regression (ROR) and using Stochastic Multiobjective Acceptability Analysis (SMAA). Finally, we shall present a didactic example that illustrates the methodology with the above variety of options. Keywords: Decision Making, Multi-Criteria Hierarchy Process, Outranking Methods. 272 MO3D (contributed) Multiple Criteria Methods in Decision Support Systems for Forest Management Segura, Marina (*) Universitat Polit`ecnica de Val`encia Maroto, Concepci´on Universitat Polit`ecnica de Val`encia Ray, Duncan Centre for Human and Ecological Sciences In decision making the most appropriate methods to solve a particular problem usually depend on specific relevant characteristics. This work provides an in-depth analysis and assessment of methods used in Decision Support Systems (DSS) for sustainable forest management. The problem nature has been considered through the following dimensions: temporal scale, spatial context, spatial scale, number of objectives and decision makers or stakeholders and goods and services involved. Some of these dimensions were significantly related and demonstrated using contingency tables. The objective dimension had a significant influence on the distribution of forest problems by temporal scale, single or multiple makers/stakeholders and the types of goods and services. For example, multiple objectives were more frequently set in strategic problems than in tactical and operational issues. We also found significant relationships between the methods and problem dimensions. The results showed that three out of four forest management problems have multiple objectives, but the percentage solved by Multiple Criteria Decision Making is only 40%. Analytic Hierarchy Process (AHP), Goal Programming, Multi-Attribute Value (MAV) and Multi-Attribute Utility Theory (MAUT) are the most frequent approaches in DSS for forest management, especially the former (AHP). The analysis demonstrated that the majority of DSSs are focused on market products, alone or together with services. Globally, at least of one third of these include multiple criteria approaches including also group decision making. In this situation multiple criteria and group decision making should be developed further in DSSs to provide a higher contribution to decision making in forestry sector. Keywords: Decision Support Systems, Forest Management, Group Decisions Making, Multi-Criteria Decision Making, Multi-Objective Optimization. 273 WE2F (invited) Multivalued Heuristics for One-To-One Biobjective Shortest Path Problems Machuca, Enrique (*) Universidad de M´alaga Mandow, Lorenzo Universidad de M´alaga We consider the problem of calculating the set of all Pareto-optimal solutions to a one-toone biobjective shortest path problem. Recent analyses have revealed that the efficiency of multiobjective label-setting algorithms improves with the use of consistent informed heuristic estimates. In particular, the vector-valued precalculated heuristic proposed by Tung and Chew (1992) has been extensively tested over the NAMOA* algorithm. The Tung-Chew heuristic is by definition the most informed admissible heuristic that can be obtained using a single vector estimate for each node. However, in theory multivalued heuristics, that provide several vector estimates for each node, can still result in even more informed heuristics. We extend the method proposed by Tung and Chew to calculate multi-valued heuristics for biobjective problems, analyze their properties, the requirements of managing multiple heuristic estimates, and evaluate the improvements in space and time requirements that can be obtained from them. Keywords: A* Search, Artificial Intelligence, Bi-Objective Shortest Path Problems, Heuristic Search. 274 WE2D (invited) Natural Gas Pipelines: A Decision Support System to Evaluate Multidimensional Risk Alencar, Marcelo (*) Universidade Federal de Pernambuco Almeida, Adiel Universidade Federal de Pernambuco Clemente, Th´arcylla Universidade Federal de Pernambuco Krym, Eduardo Universidade Federal de Pernambuco This paper presents a Decision Support System (DSS) that is able to support evaluating risks in natural gas pipelines by taking a multicriteria decision analysis into account. The use of a multi-criteria approach to risk assessment and analysis allows different aspects directly linked with the problem to be considered. In this type of approach, there is a need to collect technical, social, and environmental data relating to the pipeline sections that are to be evaluated. In addition, factors related to the decision-making process are considered as well as human value judgments. Keywords: Decision Support Systems, Multi-Criteria Decision Analysis, Natural Gas Pipeline, Risk. 275 TH4C (contributed) New Alternative Aggregation Formulas for the Human Development Index: a Multicriteria Approach Rodr´ıguez D´ıaz, Beatriz (*) University of M´alaga Angulo-Guerrero, Mar´ıa Jos´e University of M´alaga Luque, Mariano University of Malaga P´erez-Moreno, Salvador University of M´alaga Since its publication in 1990 by the United Nations Development Programme (UNDP), the Human Development Index (HDI) has generated an extensive literature that has assessed its properties, provided numerous critiques, and proposed a number of potential improvements. In 2010, coinciding with the twentieth anniversary of the launching of the first Human Development Report (HDR), the UNDP decided to revise the HDI and introduced several major changes. Many of the problems pointed out by critics were tackled with the changes introduced into the way to calculate the new HDI, although some authors considers that drawbacks still persist, particularly, in relation to the problematic tradeoffs between the HDI’s components. In this work we propose alternative aggregation formulas for the HDI and assess the problem of substitutability between the HDI’s three components. To this end, we carry out an approach based on the double reference point methodology (aspiration-reservation). For each indicator, the value of each country is normalized by means of two reference values (aspiration and reservation values) by using an achievement scalarizing function which is piecewise linear. Aggregating the new values (values of the achievement scalarizing functions), we calculate a weak indicator that allows compensation among the various components, a strong indicator that measures the state of the worst component, allowing no compensation, and a mixed indicator, which is a linear combination of these previous ones. Subsequently, the three rankings of countries resulting from the respective indicators calculated are compared to the HDI ranking, evidencing the existence of significant tradeoffs across IDH dimensions that may distort the real measurement of progress in human development among countries. Keywords: Aspiration and Reservation Values, Human Development Index, MultiCriteria Decision Analysis, Reference Point, Substitutability between components, Synthetic indicators. 276 TU3E (invited) New Developments of Interactive Multiobjective Optimization using Dominance-Based Rough Set Approach Matarazzo, Benedetto (*) University of Catania Greco, Salvatore University of Catania Slowinski, Roman Poznan University of Technology Application of Dominance-based Rough Set Approach (DRSA) in Interactive Multiobjective Optimization (IMO) gave rise to a new method employing a logical preference model. It is composed of two alternating stages: computation stage, and dialogue stage. In the computation stage, a sample of feasible solutions is generated and presented to the Decision Maker (DM). In the dialogue stage, the DM indicates relatively good solutions in the generated sample. This binary classification of sample solutions into ‘good’ and ‘others’ is an input preference information for DRSA. DRSA is using this information to induce a preference model expressed in terms of “if. . . , then . . . ” decision rules. The DM is then asked to select one decision rule considered the most interesting, which defines new constraints to be added to the constraints of the previous problem, cutting-off non-interesting solutions from the former optimal set. A new sample of solutions is generated in the next iteration from the reduced optimal set. The interaction continues until the DM finds a satisfactory solution. DRSA permits also to describe the Pareto frontier in terms of “if. . . , then . . . ” association rules. We present further developments of this methodology related to 1) possibility of selecting more than one constraining rule in each iteration; 2) possibility of considering also rules describing relatively bad solutions; 3) possibility of using DRSA not only to describe the Pareto frontier, but also the space of decision variables. Keywords: Dominance-Based Rough Set Approach , Interactive Multi-Objective Optimization, Preference Information. 277 MO5E (award) New Directions in Robustness Analysis and Preference Modeling in Multiple Criteria Decision Aiding Kadzinski, Milosz (*) Pozna´ n University of Technology We propose a set of novel decision aiding methods based on two prevailing preference models: additive value function and outranking relation. The introduced approaches take into account preference information of a new type and conduct robustness analysis of the suggested recommendation in an innovative way. The proposed robust ordinal regression procedures account for indirect, imprecise, and incomplete preference information provided by the decision makers (DMs). This includes the traditionally used pairwise comparisons or assignment examples, which are, however, decomposed in an original way to parameters of the generalized outranking- or value-based models. Additionally, we consider other types of preference information which so far has not received due attention in Multiple Criteria Decision Making. These are rank-related requirements in the form of desired positions and scores, possibly imprecise desired class cardinalities, or real intervals for indifference and preference thresholds. When it comes to robustness analysis, we present new methods for exploiting the consequences of applying all preference models compatible with the provided preferences. In particular, we implement reasoning in terms of the necessary and possible consequences for outranking methods and group decision problems. We also propose the framework of extreme ranking analysis and procedures for selection of a representative preference model instance. In this way, the introduced approaches identify the recommendation observed in case of all, some, the most favorable, the least advantageous, or the most robust compatible model for the considered preferences. When confronted with the value system of the DM, these results can be used for looking more thoroughly into the subject, by exploring, reasoning, or testing scenarios. Thus, the presented methods promote alternate phases of preference elicitation and robustness analysis as a versatile tool for approaching real-world decision problems. Keywords: Additive Value Function, Decision Aiding, Outranking Relation, Preference Disaggregation, Preference Modeling, Robust Ordinal Regression, Robustness Analysis. 278 MO3E (contributed) NIMBUS in Interactive Multiobjective Optimization under Uncertainty Miettinen, Kaisa (*) University of Jyvaskyla Mustajoki, Jyri Finnish Environment Institute Stewart, Theodor University of Cape Town We propose an interactive method for decision making under uncertainty, where uncertainty is related to the lack of understanding about consequences of actions. Such situations are typical especially in design problems, where a decision maker has to make a decision about a design at a certain moment of time even though the actual consequences of this decision can be possibly seen only many years later. To overcome the difficulty of predicting future events when no probabilities of events are available, our method utilizes groupings of objectives or scenarios to capture different types of future events. Each scenario is modelled as a multiobjective optimization problem to represent different and conflicting objectives associated with the scenarios. We utilize the interactive classification-based multiobjective optimization method NIMBUS for assessing the relative optimality of the current solution in different scenarios. This information can be utilized when considering the next step of the overall solution process. Decision making is performed by giving special attention to individual scenarios. We demonstrate our method with an example in portfolio optimization. Keywords: Classification, Decision Making, Interactive Methods, Multi-Objective Optimization, NIMBUS, Pareto Optimal Solutions, Scenarios, Uncertainty Handling. 279 TU2E (invited) Non Additive Robust Ordinal Regression for Urban and Territorial Planning: an Application for Siting an Urban Waste Landfill Bottero, Marta (*) Ferretti, Valentina Politecnico di Torino Angilella, Silvia Lami, Isabella Corrente, Salvatore Greco, Salvatore University of Catania Multiple Criteria Decision Aiding (MCDA) gives not only a toolbox, but, overall, a well developed methodology to support decision-making processes. MCDA is particularly useful in the context of sustainability assessment and urban and territorial planning, where a complex and inter-connected range of environmental, social and economic issues must be taken into consideration. Moreover, the analysis of the possible interactions among the considered criteria is of particular importance for assessing the sustainability of a certain transformation. The paper proposes an application of the Non Additive Robust Ordinal Regression approach for the selection of a new landfill location among several alternatives. NAROR aggregates information on the considered criteria using the Choquet integral in order to represent the interactions among criteria by means of non-additive weights, technically called fuzzy measures. Using NAROR, the DM can supply preference information in terms of pairwise preference comparisons, intensity of preferences, pairwise comparisons on the importance of criteria and sign and intensity of interaction among pairs of criteria. The basic idea of NAROR is to consider the whole set of the fuzzy measures that are compatible with the preference information given by the DM. In fact, the recommendation supplied by NAROR is expressed in terms of necessary preferences, in case an alternative is preferred to another for all compatible non-additive weight vectors, and of possible preferences, in case an alternative is preferred to another for at least one compatible non-additive weight vector. In the considered case study, several sites for the location of the landfill are analyzed and compared through the use of the NAROR on the basis of different criteria, such as presence of population, hydrogeological risk, interferences on transport infrastructures and economic cost. Keywords: Fuzzy Measures , Interaction Among Criteria, Multi-Criteria Decision Aid, Sustainability Assessment. 280 TU3E (invited) Non-Linear Multi-Objective Optimization via Parallel Algorithms: Solving a Planar Facility Location and Design Problem L´opez-Redondo, Juana (*) University of Granada Fern´andez-Hern´andez, Jos´e University of Murcia Gila-Arrondo, Ar´anzazu University of Murcia Mart´ınez-Ortigosa, Pilar University of Almer´ıa A new multi-objective evolutionary algorithm, called FEMOEA, which can be applied to many nonlinear multi-objective optimization problems has been recently proposed. It combines ideas from different multi- and single-objective optimization evolutionary algorithms, although it also incorporates new devices, namely, a new method to improve the efficiency of points using gradient information and a new stopping rule, which help to improve the quality of the obtained approximation of the Pareto-front and to reduce the computational requirements. Even so, when the set approximating the Pareto-front must have many points (because a high precision is required), the computational time needed by FEMOEA may be not negligible at all. Furthermore, the computational resources needed may be so high that a PC may run out of memory. In those cases, parallelizing the algorithm and run it in a supercomputer may be the best way forward. In this work, a parallelization of FEMOEA, called FEMOEA-Paral, is presented. To show its applicability, a bi-objective competitive facility location and design problem is solved. As the results will show, FEMOEA-Paral is able to maintain the effectiveness of the sequential version and this by highly reducing the computational costs. In fact, super efficiency values were obtained independently of the number of available processors. Furthermore, the parallel version shows a good scalability. Keywords: Bi-Objective Location Problem, Effectiveness, Efficiency, Non-Linear MultiObjective Optimization, Parallel Algorithm. 281 MO4B (invited) Nonconvex Pareto Navigator for Interactive Multiobjective Optimization Hartikainen, Markus (*) University of Jyv¨askyl¨a Klamroth, Kathrin University of Wuppertal Miettinen, Kaisa University of Jyvaskyla This talk describes a new interactive method called Nonconvex Pareto Navigator which extends the convex Pareto Navigator method for nonconvex multiobjective optimization problems. The Nonconvex Pareto Navigator method enables moving on an approximation of the Pareto front in the objective space so that the decision maker can find preferable compromise values for the objectives. Navigating on the approximation allows fast computation and, thus, a better user experience for the decision maker, when compared to navigating on the actual Pareto front. The method is especially useful in solving computationally expensive real-life multiobjective optimization problems where function evaluations may be time-consuming. The proposed Nonconvex Pareto Navigator method consists of two essential parts: constructing a Pareto front approximation and navigating on it. In this talk, besides introducing the new method, we discuss quality criteria - both mathematical and those related to the decision maker’s user experience – which the approximation and the navigation should fulfill. Finally, we demonstrate how our proposed Nonconvex Pareto Navigator performs with respect to those criteria. Keywords: Decision Making, Interactive Methods, Multi-Objective Optimization. 282 WE1C (contributed) On a Linear Transformation of Variables in Data Envelopment Analysis Korhonen, Pekka (*) Aalto University Keshvari, Abolfazl Aalto University In this presentation, we consider a linear transformation of the input- and output-variables in the Data Envelopment Analysis. When a linear transformation is non-singular, no problems arise provided the efficient frontier is re-defined, respectively. Instead, a singular linear transformation is much more complex. This is the case, when we reduce the number of inputs and/or outputs. We will consider two dimension reduction methods commonly applied in the Data Envelopment Analysis (DEA) and point out that a user has to be very careful in choosing the method. In reduction of the number of output/input-variables, there are several pitfalls, which may lead to wrong results. We will discuss these pitfalls and illustrate problems with numerical examples. Keywords: Data Envelopment Analysis, Linear Transformation, Variable Reduction. 283 TU3C (contributed) On Application of Multi-Criteria Decision Making with Ordinal Information in Elementary Education Mazurek, Jiˇr´ı (*) Silesian University in Opava In the Czech Republic each elementary or secondary school decides which textbook will be used for a given class and a given subject of education. As a supply of textbooks is wide, a selection of the most suitable textbook by a teacher is a typical case of multi-criteria decision making situation where an evaluation of different textbooks on selected criteria is rather ordinal in nature than cardinal, as it is not possible to assign textbooks some numerical value with regard to criteria such as content, comprehensibility, adequacy to children’s age and knowledge, etc. (with the exception of textbook’s price), but textbooks can be ranked from the best to the worst by such criteria, and the best textbook can be found by a new and simple mathematical method developed for this purpose in this paper. The aim of the paper is to show how this multi-criteria decision making method with ordinal information can be used for the selection of the most appropriate textbook for elementary science education, because a right choice of a textbook plays an important role in children’s education. And we shall not forget that decisions made today influence the world tomorrow, and the World of Tomorrow is also a World of Our (well-educated) Children. Keywords: Elementry Education, Multi-Criteria Decision Making, Ordinal Information. 284 FR2E (invited) On Approximate Kernels of Minimal Size for Bicriteria Problems Jamain, Florian (*) LAMSADE Bazgan, Cristina LAMSADE Vanderpooten, Daniel LAMSADE Most multicriteria combinatorial optimization problems admit families of instances for which the number of non-dominated points is exponential in the size of the instance. For computational, as well as cognitive reasons, it is often relevant to generate a reduced subset of solutions that represents as well as possible the different choices, i.e. providing a good approximation of the Pareto set. This idea is represented by the concept of an ePareto set, which is a set of solutions that approximately dominates every other solutions, i.e. such that for every solution S it contains a solution S’ that is better within a factor (1+e) than S in all the objectives. Note that there may exist several e-Pareto sets of very different size. An interesting problem studied by Vassilvitski and Yannakakis (2005) and Diakonikolas and Yannakakis (2009) is the efficient construction of an e-Pareto set of size as small as possible. In spite of this, such small e-Pareto sets may contain pairs of solutions such that one of them e-dominates the other one. Therefore, it may be interesting to impose as an additional constraint that an e-Pareto set should only contain solutions which are pairwise incomparable relatively to the e-dominance relation. Such e-Pareto sets are called e-kernels. In this work, we propose generic algorithms that compute in polynomial time an e-kernel of size as small as possible. We show some general results about e-kernels and present a polynomial time algorithm to construct small e-kernels under some conditions for the bicriteria case. We show that an e-kernel does not always exist when the number of criteria is at least three. Keywords: Approximation Algorithm, Minimal Size, Multi-Criteria Decision Model, Pareto Set. 285 TH4D (contributed) On Branch and Bound Approach to Multi-Objective Global Optimization Zilinskas, Antanas (*) Vilnius University Zilinskas, Julius Vilnius University The first part of the talk is on the worst case optimal algorithms for Lipshitzian objectives. Nonconvex multi-objective optimization problems are considered. A result of single objective global optimization is generalized showing that the worst case optimal algorithm for approximation of Pareto front is reducible to the covering the feasible region by balls of minimum radius. The adaptive and passive algorithms are coincident. A version of onestep optimal algorithm is analyzed in detail and implemented using the branch-and-bound approach. Numerical examples are provided to compare the developed algorithm with an algorithm optimal in average, and with some metaheuristics. The second part of the talk concerns practical applications of the algorithms based on branch-and-bound approach. Two combinatorial problems are considered: machine grouping for manufacturing, and drawing aesthetically appealing graphs. Cell formation problem aims at identification of groups of machines to form manufacturing cells. The problem involves conflicting objectives, for example, the total intercell moves of parts are minimal when all machines are in a single cell, but the within-cell load variation is minimal when each machine composes a separate cell. In this paper we propose a bi-criterion branch and bound algorithm for cell formation problem. Two bi-criterion models were used: minimization of the numbers of exceptional elements and voids, and minimization of intercell moves and within-cell load variation. We demonstrate the performance of the algorithm solving problems from the literature and compare the results with those of metaheuristic algorithms from the literature. The second applied problem is related to the drawing of special graphs related to business process management. The problem is reformulated as a multi-objective optimization problem where the quantitatively measurable aesthetic criteria are involved. An algorithm is justified and numerical examples are provided. Keywords: Approximation, Branch and Bound, Combinatorial Optimization, MultiObjective Optimization. 286 TH2F (contributed) On Continuation Methods for Non-Linear Multi-Objective Optimization Martin, Benjamin (*) University of Nantes Goldsztejn, Alexandre University of Nantes Granvilliers, Laurent University of Nantes Jermann, Christophe University of Nantes Recent advances in Non-Linear Multi-Objective Optimization (NLMOO) have focused on the use of continuation methods that exploit the continuity of solutions satisfying the first order optimality conditions. Given a single solution, continuation methods allow the inexpensive computation of a full component of connected solutions. The growing interest in combining continuation with NLMOO methods is noticeable in the recent literature. For instance, in Sch¨ utze et al. (2008) and Harada et al. (2007), continuation is associated with metaheuristics (respectively a Particle Swarm Optimizer and a Genetic Algorithm) in order to build solutions well scattered upon the Pareto set. With a similar goal, Pereyra et al. (2009-2011) propose a continuation coupled with global optimization techniques in the bi-objective case. Within a set-oriented framework, Sch¨ utze et al. (2005) use continuation to fasten the computation of new portions of the Pareto set, and to recover connected parts possibly missed by the algorithm. Lovison (2011-2012) proposes a deterministic approach, using Delaunay tessellations and continuation to cover sets of solutions satisfying optimality conditions. This approach is global provided the tessellation cuts every connected components of solutions. We propose a comparative overview of such NLMOO methods using continuations with the objective of showing their potential and weaknesses. One discussion of special interest is their ability to handle constrained NLMOO. We also introduce a novel rigorous continuation method based on recent advances in Constraint Programming. Used for NLMOO, it can in principle reconstruct a certified enclosure of the whole Pareto-optimal set provided one solution in each component of this set is given, i.e., under assumptions similar to Lovison (2011). Finally, we discuss how such initial guesses could be obtained. Keywords: Constraint Programming, Hybrid Approaches, Multi-Objective Optimization, Non-Linear Multi-Objective Optimization, Path-Following Methods. 287 WE1D (invited) On the Need for Multi-Criteria Modelling in Power Systems Analysis and Planning Bertsch, Valentin (*) Karlsruhe Institute of Technology Fichtner, Wolf Karlsruhe Institute of Technology The promotion of renewable energy sources (RES) and combined heat and power generation leads to an increasing decentralisation of energy systems and brings about new challenges. Especially in Germany, the realisation of the ambitious targets concerning the expansion of RES necessitates an extensive structural rearrangement of the power system. For instance, the number of players in the market recently increased in consequence of the decentralisation and is expected to further increase. Since the different players typically pursue different objectives and have different preference perceptions, multiple and usually conflicting targets need to be considered. As a result, decision processes need to be designed in a participatory way. Moreover, a purely economic optimisation is no longer sufficient to support decision making in energy systems since the importance of ecological, technical and socio-psychological criteria steadily increases. We therefore call for combining multi-criteria decision analysis (MCDA) and energy systems modelling approaches. Because of its transparent nature, multi-attribute value theory (MAVT) seems to be suitable to bring together quantitative (e.g. economic, ecological) and qualitative (e.g. acceptance) information in such participatory decision processes. Energy systems models can be used to generate quantitative input data for the multi-criteria analysis. Therefore, a multi-period linear optimisation model is presented, which provides a regional capacity expansion planning approach. The model is applied in a scenario-based analysis in order to analyse the impact of the uncertainties associated with the long-term development of the power system. Moreover, sensitivity analysis techniques are presented, which may facilitate the process of preference elicitation by exploring the robustness of the analysis with respect to variations of the subjective parameters. Finally, possible further developments of model and methodology are discussed. Keywords: Decision Maker’s Preferences, Linear Programming, Multi-Attribute Value Theory, Participatory Multi-Criteria Evaluation, Sensitivity Analysis, Uncertainty Handling. 288 MO2B (contributed) On the Optimal Inputs Balance in Terms of the Educational Multiobjective Model: the Case of Primary and Secondary Education Students Luque, Mariano (*) University of Malaga L´opez Agudo, Luis Alejandro University of M´alaga ´ Marcenaro-Guti´errez, Oscar University of M´alaga This paper sheds new light on the relationship between inputs and outputs in the framework of the educational production function. In particular it is intended to gain a better understanding of which factors may be affected to achieve an optimal educational outputs level. To this aim we analyze teacher and test-based assessments in particular subjects using a multiobjective schema. For much of the analysis we use data from a very recent (2010) Survey (focused on Education and Housing in Andalusia) linked, on one hand, with the results from an educational assessment program conducted among 10 and 14 year students (containing information on a wide set of per-sonal, family and school environment characteristics) and, on the other hand, with the adminis-trative records on teacher-based scores. These data are used to build up a mixed integer multiobjective model, on the basis of a previous statistical and econometric analysis of these data. Then, a method based on a reference point approach is used to determine the profile of the most ’successfully -balanced’ students in terms of educational outcomes. Using appropriate reference levels and weights, reference point methodology allow to generate “very balanced” solutions in terms of the objective function values (educational outcomes). Finally, a sensitive analysis is used to determine policies than can be carried out in order to increase the performance levels of primary and secondary education students. The paper thus offers a new and crucial contribution to the debate on education policies, particularly relevant in a context of economic crisis. Keywords: Econometrics, Education, Multi-Objective Optimization, Reference Point. 289 MO2B (contributed) Optimization of Investment in Shares by Multi-Objective Optimization Brauers, Willem K. M. (*) Vilnius Gediminas Technical University Is it possible to use MOO for the selection of shares on a Stock Exchange such as shares of the DJI New York, FTSE100 London, Nikkei225 Tokyo, CAC40 Paris etc? The as best selected shares would be the outcome of eight objectives. Weights in many MO-methods have a double function: normalization of the different units and giving importance to the objectives. By obtaining dimensionless measures through the reading of the matrix in a vertical way normalization is not needed. Up till now all objectives were based on equal importance. The introduction of the opinions of the analysts will give more importance to some objectives. For the FTSE100 the matrix is composed of 100 shares to 9 objectives or in total 900 elements. Two different methods under the name of ’MOORA’, namely a Ratio System and a Reference Point Approach, will try to make optimal the content of the matrix. By adding the Full Multiplicative Form, three methods based on dimensionless measures will be used. They will compose ’MULTIMOORA’, whereas the three methods will control each other. An Ordinal Dominance Theory ranks the outcomes of the three methods. The Belgian Bell20 Index is used as an example as it presents a smaller exercise without too many difficulties to find the necessary data. However the composition of the Belgian Index changes regularly making comparisons over time much more difficult. In addition, the financial importance of Belgium diminished considerably. Some General remarks: 1) the application concerns the past. Pure extrapolation has no sense for such a fluctuating market. Regularly revisions are needed 2) companies with an effective management are assumed. A thorough study work on each company is necessary 3) there are the Unknown Unknowns or may we say the Economics of Uncertainty? This investigation is just an exercise, but with the aim how to operate similar studies. Anyhow it has to be stressed that never before such a study for another World Index on shares occurred. Keywords: Finance, Multi-Objective Optimization, Multimoora, Multiplicative Form, Ratio Analysis, Reference Point. 290 MO3F (contributed) Optimization of the Operation of the Auxiliary Services of Power Plants using Preference-based Evolutionary Optimization Ruiz, Ana Bel´en (*) University of M´alaga Cabello, Jose Manuel University of M´alaga Luque, Mariano University of Malaga Ruiz, Francisco University of M´alaga The auxiliary services of a power plant are required for the plant operation. While they are not the main components, their energy consumption is often significant, and it can be reduced by implementing a series of strategies, like installing capacitors in different parts of the network, changing current motors by high efficiency ones and/or installing variable speed drives. Carrying out these improvements has a direct impact on the electricity consumption of the auxiliary systems and, thus, on the electricity required for the plant operation. Overall, the interest of these energy savings lies, on the one hand, on the reduction of the electricity production costs and, on the other hand, on the decrease of the CO2 emissions. However, the cost of implementing these changes can be very high, and it has to be evaluated. Moreover, a further economic analysis should be considered in order to maximize the profitability of the investment. Therefore, we have proposed a multiobjective problem that determines which strategies are most suitable to maximize the energy saving, to minimize the economic investment and to maximize the Internal Rate of Return of the investment. In order to solve this multiobjective problem together with a decision maker (DM), we have designed a process which progressively interacts with him/her in order to, firstly, describe the auxiliary services to be studied and, secondly, to find the most preferred solution according to his/her preferences, which are indicated by means of desirable objective function values (a reference point (RP)). In that resolution process, the evolutionary approach NSGA-II is applied to analyze the trade-offs among the objectives, and, subsequently, the preference-based evolutionary algorithm WASF-GA is used to approximate the reachable area of the nondominated objective set from the RP. Finally, we have solved the problem associated to the auxiliary services of a particular case study together with a real DM. Keywords: Auxiliary Services Of Power Plants, Multi-Objective Optimization, PreferenceBased Evolutionary Algorithms, Reference Point. 291 TH2F (contributed) Optimization over the Weakly Efficient Set of a Multiobjective Programming Problem Constrained by Quadratic Functions Yamada, Syuuji (*) Niigata University Tanaka, Tamaki Niigata University Tanino, Tetsuzo Osaka University In this talk, we consider a multiobjective programming problem (MOP), where all objective functions are linear and all constrained functions are quadratic. It is known that many multiobjective nonlinear programming problems can be transformed into or approximated by (MOP). If (MOP) is constrained by a single quadratic function, the weakly efficient set can be defined as the intersection of a closed convex cone and the boundary of the feasible set. Based on this fact, we formulate a quadratic programming problem (QP) for checking the weak efficiency of a given feasible solution in the case where (MOP) is constrained by multiple quadratic functions. Moreover, it is shown that (QP) can be solved by listing Karush-Kuhn-Tucker points. By incorporating such the techniques into the standard nonlinear optimization algorithms, we propose a globally optimization algorithm for a problem (MP) to minimize another linear function over the weakly efficient set. An example of (MP) is furnished by the portfolio optimization problem in capital markets. A fund manager may look for a portfolio which minimizes the transaction cost on the efficient set. In order to verify the effectiveness of the proposed algorithm, we show the computational experiments. Keywords: Global Optimization, Multi-Objective Optimization. 292 TU5C (contributed) Optimizing Economic and Environmental Performances of Solar Power and Electric Vehicles: A MOMILP Application De Schepper, Ellen (*) Lizin, Sebastien Hasselt University Gandibleux, Xavier Van Passel, Steven Martin, Benjamin Vincent, Thomas University of Nantes The EU has set targets for reducing its greenhouse gas emissions progressively up to 2050. Recognizing that the sectors of heat & electricity generation and transport are the world’s largest contributors to climate change, the deployment of clean energy and transportation technologies is widely stimulated. Unfortunately, better environmental performances often imply higher economic costs. This trade-off between two conflicting objectives calls for a multi-objective optimization (MOO) assessment, aiming to find the “best” possible solutions, i.e. the Pareto optimal solutions. Whereas the economic and ecological optimization of energy systems is extensively studied in literature, little research has been done on transportation systems. Furthermore, we argue that it is valuable to simultaneously optimize energy and transportation systems for two reasons. First, most entities (e.g. firms, areas, individuals) have needs regarding both energy and transportation. Second, when considered simultaneously, synergies between the energy and transportation systems can be exploited. This paper aims at filling this gap by performing a MOO on a Belgian case study, i.e. a SME having a certain need for electricity and traveling. The considered energy technologies are solar photovoltaics and grid electricity; for transport internal combustion engine vehicles, grid powered battery electric vehicles (BEVs), and solar powered BEVs are available. Aiming to obtain realistic results, the possible existence of scale economies is taken into account. The latter implies a mixed integer programming problem. Accordingly, this paper applies the exact algorithm described in: T. Vincent, F. Seipp, S. Ruzika, A. Przybylski, X. Gandibleux; Multiple objective branch and bound for mixed 0-1 linear programming: Corrections and improvements for the biobjective case. Computers & Operations Research, 40(1)498-509, 2013. Additionally, the impact of policy measures on the Pareto front is visualized. Keywords: Case Example, Mixed Integer Programming. 293 TU4F (contributed) Optimizing Multi-objective Mixed Integer Programs using Evolutionary Algorithms Stidsen, Thomas (*) Technical University of Denmark Solving Multi-Objective Mixed Integer Program (MOMIP) models is notoriously difficult. The standard techniques are the Two-Phase method or the Epsilon method, which both rely on solution of Single Objective MIP’s. But if the solution of the single-objective MIP is too hard, both methods fail. On of the more successfull approaches to solve MultiObjective optimization problems has been Evolutionary Algorithms like e.g. NSGA. Traditionally meta-heuristic algorithms has had problems solving Mixed Integer Programs, primarily because it is hard to ensure feasibility of the solutions, given non-trivial constraints. Just generating feasible start solutions is (NP) hard. In this presentation I will present a novel approach where an Evolutionary Algorithm is used, in combination whith a standard MIP solver, to solve heuristically, the MOMIP, finding a heuristic Pareto front. By using the MIP solver, the traditional problems with feasibility are avoided. A standard MIP solver is used for crossover and mutation operators and any of the well-known evolutionar algorithm schemes, e.g. NSGA can be used for selection. Keywords: Evolutionary Computations, Evolutionary Multi-Objective Optimization, Integer Programming, Mixed Integer Programming, Multi-Objective Optimization. 294 WE2E (contributed) Optimizing Telecommunications in Vehicular Networks with a Parallel Multiobjective PSO Toutouh, Jamal (*) University of M´alaga Alba, Enrique University of Malaga This work presents pSMPSO, a master/slave version of the Speed-constrained Multiobjective Particle Swarm Optimization (SMPSO) algorithm that runs mutation and evaluation in parallel to accelerate the search. Our pSMPSO proposal has been validated on a real world problem in smart cities (not on an academic benchmark): the optimization of vehicular ad hoc network (VANETs) communications. We make our algorithm to find an efficient parameter setting of the Ad hoc On Demand Distance Vector (AODV) routing protocol. Our pSMPSO will optimize two conflicting quality-of-service goals: maximize the packet delivery ratio (PDR) and minimize the end-to-end delay (E2ED). The evaluation of every solution requires its simulation with real maps and traffic, what needs long running times (around 46 s.). The experimental validation of the proposed pSMPSO is performed by analyzing 15 independent runs of two multi-thread variants named pSMPSO-8 and pSMPSO-16. They are implementations with 8/16 particles in the swarm (8/16 threads). The average run time of executing pSMPSO-8 and pSMPSO16 in a single processor is 7,773.62 and 11,603.73 m., respectively. Their multithreaded versions on a parallel computer reduce those times to 578.76 and 584.87 m., respectively. Thus, the average speedup is 6.88 for pSMPSO-8 and 13.43 for pSMPSO-16, showing a desirable almost-linear speedup. Also, the average computational efficiency is 85.96% (8 threads) and 83.95% ( 16 threads), really high values. In order to ease decision making we have selected representative solutions of the Pareto front and analyzed the induced behavior in communications of the real scenario. Compared to the standard RFC 3561 and to the optimized configuration got by a sequential and parallel PSO we can confirm that our pSMPSO outperforms the rest in terms of PDR and E2ED. Moreover, these improvements increase with the number of vehicles and the communications needs, i.e. they do scale to real cities. Keywords: Data Routing, Parallel Multi-Objective Optimization, Software Protocol, Vehicular Ad Hoc Networks (VANET). 295 WE2G (invited) Pareto Equilibria in Cournot Competition Under Uncertainty Caraballo, M. Angeles (*) University of Seville M´armol, Amparo Mar´ıa University of Seville Buitrago, Eva University of Seville Monroy, Luisa University of Seville In this paper we investigate a competitive situation in which two firms producing a homogeneous commodity have to decide the quantity to offer. There is uncertainty on the price the commodity will reach in the market since different future scenarios are possible and it is assumed that no information about the probability distribution of occurrence of the scenarios is available. This decision-making situation is formalized as a game with incomplete preferences represented by vector-valued utility functions for which the notion of Pareto equilibrium is adopted as a natural extension of that of Cournot equilibrium. Under standard assumptions about the demand curve, we characterize the complete set of Pareto equilibria. In the second part of the paper we analyse the equilibria to which the agents will arrive depending on their risk attitudes. We find that equilibria always exist if both agents are simultaneously either optimistic or pessimistic. However, this is not the case when each of the agents exhibits a different and extreme attitude towards risk. Keywords: Cournot Competition, Pareto Equilibria, Strategic Game. 296 WE2D (invited) Plaster Waste Destination Problem in the Buildings Sites: An Application of the VFT Methodology Hazin, Luciana (*) Universidade Federal de Pernambuco Alencar, Marcelo Universidade Federal de Pernambuco Mota, Caroline Universidade Federal de Pernambuco It is each time greater the interest to find solutions that attends the interested parties and, at the same time, that decrease the impact in the environment originated by the plaster wasted, that is a material which is widely used in Brazil on construction sites. With the view to the growing concern with the environment, it becomes each time more important the understanding of the need of reducing the amount of residue generated in new works, reducing the losses, and of educating the society in the sense of recycling the dump generated in the reforms and demolitions. In this sense, the value focus thinking methodology (VFT), that is respect to how the values can be used to improve the decision making process, was applied. The problem was structured and the parties involved could get a better understanding regarding their objectives and the consequences for the decision problem. Keywords: Plaster Waste, Structuring Decisions, Value Focus Thinking. 297 TU4G (contributed) Portfolio Multiple Criteria Analysis, Ranking and Optimal Selection Fulga, Cristinca (*) Bucharest University of Economic Studies In this paper we develop a portfolio selection method in the Mean-Risk framework. The recent positive evolution of the risky assets that are not comprised in the portfolio available at the moment of the decision is analyzed by using the Analytical Hierarchy Process. Our model relies on a quantile based risk measure defined using the modified loss distribution according to the decision maker’s risk and loss aversion preferences. We investigate the practical performance of the proposed model on a portfolio composed of some of the most representative securities of the New York Stock Exchange. Keywords: Multi-Criteria Decision Making, Portfolio Optimization. 298 TU4G (invited) Practical Measurement of Environmental and Social Achievement Levels for Mutual Funds: A Synthetic Indicator with its Software Application Pla Santamaria, David (*) Bravo Sell´es, Milagros Garc´ıa Bernabeu, Ana M. Universidad Polit´ecnica de Valencia P´erez-Gladish, Blanca University of Oviedo This paper has the following purposes: (i) to measure in an objective way the environmental and social levels of achievement for financial assets and mutual funds from multiple criteria of socially responsible investment; (ii) to provide a software application, which allows the fund managers to easily implement the proposed ethical measurement. Purpose (i) requires constructing a synthetic indicator, which does not depend on particular preferences of fund managers, since ethical preferences of investors in mutual funds widely change from an investor to another. Accordingly, the synthetic indicator is constructed from domination analysis and decision making analysis under uncertainty. A veto against the dominated alternatives is proposed. Purpose (ii) requires constructing a specific tool (the software application), whose framework and practical use are shown. An actual case from a large number of criteria is developed and commented. Keywords: Environmental Decision Making, Equity Mutual Funds, Social Decision Making, Socially Responsible Investment, Synthetic indicators. 299 MO5H (poster) Precise Consistency Consensus Matrix in an AHP-Group Decision Making Local Context Moreno-Jim´enez, Jos´e Mar´ıa (*) University of Zaragoza Aguar´on, Juan University of Zaragoza Escobar, Mar´ıa Teresa University of Zaragoza A fundamental factor in multiactor decision making is the possibility of reaching consensus solutions among the actors involved in the resolution process. In general, the consensus may refer to several aspects of this process, such as, for example, the approach or technique that is followed, the methodology employed and the procedures used when aggregating preferences, priorities and judgments. Assuming a multiactor local context (a unique criterion), analysed with the Analytic Hierarchy Process (AHP) and using the row geometric mean as the prioritisation procedure, this paper presents a new decisional tool based on the property of consistency. It improves the behaviour of a previously defined tool (the Consistency Consensus Matrix - CCM) for AHP-Group Decision Making. This new tool, the Precise Consistency Consensus Matrix (PCCM), allows an automatic increase in the number of entries in the CCM by identifying precise values in common consistency intervals. The PCCM results in more informed group decision making and provides more accurate estimations for the group’s priorities. It can also be used as a starting point for posterior negotiation processes. The new decisional tool has been applied to a public decision making procedure. Keywords: Analytic Hierarchy Process, Consistency, Consistency Consensus Matrix, Group Decisions Making. 300 TU5D (invited) Price-setting Under Risk Criteria for Virtual Products and Services Henig, Mordecai (*) Tel Aviv University Avinadav, Tal Bar Ilan University Chernonog, Tatyana Bar Ilan University This work deals with pricing of “virtual” products, i.e., products that a retailer can supply after demand has been realized. Such products allow the retailer to avoid holding costs and ensure timely fulfillment of demand with no risk of shortage. Assuming that demand is price-dependent and uncertain, we seek to optimize each of three criteria: expected profit, the likelihood of achieving a profit target, and a profit percentile criterion. Simultaneous multiple criteria are also explored. Two cases of demand uncertainty forms are analyzed: the multiplicative form, where a stochastic dominance is proved so that all the criteria are optimized simultaneously; and the additive form, where stochastic dominance cannot occur. A two-echelon supply chain (comprising both the supplier and the retailer) is analyzed to show that in the multiplicative form of demand, stochastic dominance occurs in both echelons. Keywords: Multi-Criteria Decision Analysis, Pricing, Revenue Management, Risk Analysis, Supply Chain Management. 301 TH3D (contributed) Primitive and Derived Information in Decision Models Tsouki`as, Alexis (*) LAMSADE - CNRS Colorni, Alberto Politecnico di Milano The paper presents a general framework describing decision models. We specially focus on the so-called primitives of the model, that is information that cannot be derived, but needs to be learned from preference statements provided by the decision maker. We classify preference statements and what can be induced from them. We then show what derived information can be constructed. The aim of this framework is on the one hand to show the limited number of independent methods (in the sense of independent characterising primitives) and on the other hand to provide the basis for justifications and explanations to be added to decision support methods. Keywords: Decision Analysis, Decision Support, Preference Learning. 302 TU3B (invited) Prioritizing Third Party Logistics Providers: An Analytical Network Process Approach Karpak, Birsen (*) Youngstown State University Bayazit, Ozden Central Washington University As shippers worldwide are increasing their use of third party logistics (3PL) services, the need for carefully selecting 3PLs is becoming much more critical. Choosing the right 3PL provider is often a dilemma faced by supply chain managers. Our purpose is to show how the Analytic Network Process (ANP) can be used in aiding companies with the decision of selecting the most capable third party logistics provider for an aerospace company. The aim of this research is to investigate a case of 3PL provider selection decision analysis using ANP along with a sensitivity analysis to demonstrate the effectiveness of the proposed approach. Data for the study will be collected from an aerospace company located in Western Washington. We are hoping that the model we will develop in this study will help companies evaluate outsourcing logistics services. Keywords: Analytic Network Process, Third Party Logistic Supplier. 303 MO2D (contributed) Problems, Objectives and Ecosystem Services in Protected Areas: A Mediterranean Case Study Segura, Marina (*) de Castro Pardo, M´onica Maroto, Concepci´on Ginestar, Concepcion Universitat Polit`ecnica de Val`encia Management of protected areas should mainly be addressed using Multiple Criteria and Group Decision Making approaches. We need to take into account the multiple conflicting objectives and the preferences of all of the stakeholders involved. The first step when applying MCDM to the management of protected areas is to identify the decision makers and the stakeholders, the main strategic problems and objectives, as well as their priorities for each group of stakeholders. The objective of this work is to elicit and analyse the preferences of stakeholders about problems, objectives and ecosystem services in the network of natural parks in the Comunitat Valenciana, Spain. We have designed and carried out a survey that has been sent to all 22 Valencian natural parks. Firstly, we have analysed the global data to define a general framework. The natural parks in this Mediterranean area can be classified as protected wetlands, forest and marine protected areas. The nature of problems and conflicts among stakeholders are very different in each category. Secondly, we focused on two particular parks. One of them is the biggest forest natural park in the Comunitat Valenciana,Sierra de Espad´an, characterized by low population density and where municipalities are important owners. The second park is a wetland in a metropolitan area,l’Albufera de Valencia, with high population density and a lot of groups of stakeholders with a high level of conflict. In this case we have obtained responses from the park director, staff, and the representatives of municipalities, private owners, farmers, fishers, hunters, unions, conservation associations and others. We have also carried out a survey focused on this wetland using the social networks Facebook and Twitter. Finally, a discussion of results and conclusions will be presented. This work represents the first phase in implementing a collaborative and multiple decision making process to solve specific problems in these protected area. Keywords: Case Study, Decision Maker’s Preferences, Ecosystems, Group Decisions Making, Identifying Objectives, Multi-Stakeholder Decision Making. 304 WE1C (contributed) Production Points with the Same Reference Hyperplane in DEA: Applications in Sensitivity Analysis and Returns to Scale Classification Dehnokhalaji, Akram (*) Aalto University Nasrabadi, Nasim Birjand University Soleimani-Damaneh, Majid University of Tehran & IPM In this study, we consider a given supporting hyperplane of the Production Possibility Set (PPS) denoted by H and obtain a subset of the PPS with the property that for each production point belonging to that, the radial projection point lies on H when we assess the production activities by CCR or BCC model in DEA. The results are interesting not only from the theoretical viewpoint but also from the applied point of view. Theoretically, we have established some lemmas, theorems and corollaries using linear programing and DEA techniques to construct a mathematical formulation for the above mentioned subset. Identifying a set of production activities having the same rate of trade-offs is one application of our study. Also, the theory enables the decision maker to find a preservation region for preserving the Returns to Scale Classification of the DMUs. Keywords: Data Envelopment Analysis, Preservation Region, Radial Projection, Reference Hyperplane, Returns To Sclae, Sensitivity Analysis. 305 WE1E (invited) Promoting Sustainability in Mining: Multicriteria Assignment of Resources in Vale Autran M. Gomes, Luiz F. (*) Ibmec/RJ Camanho, Roberto Tipec/SP Macedo, Marcelo G. C. Vale/Toronto Camanho, Thomaz M. AMPLIX - Consultoria em Excel Martins, Luiz Geraldo Biagioni Ibmec/RJ The mining and metals industry is core to the global economy. Sustainable development is a multidimensional concept, including socio-economic, ecological, technical and ethical perspectives. As a consequence, sustainability issues are characterized by a high degree of conflict. Mining companies, like all businesses, focus on financial viability and profitability. The mining sector’s environmental and social footprint can be significant, with impacts at local, regional, national and global levels. MCDM is a significant methodology to aid in getting a comprehensive understanding about the sustainability performance of mining companies. In this paper we consider as a realistic example the case of the Brazilian mining company Vale and show how sustainability issues can be tackle in this company by using the AHP coupled to a linear program for assigning resources. We close the paper with general guidelines for choosing MCDM analytical methods for promoting sustainability in the mining sector. Keywords: Allocation, Analytic Hierarchy Process, Cost Benefit Analysis, Linear Programming, Sustainability. 306 TH2E (invited) Properties and Benefit of the Linear Efficient Frontier Approximation in the Objective Space Maag, Volker (*) Fraunhofer ITWM Grebe, Tabea Fraunhofer ITWM Nowak, Uwe Fraunhofer ITWM Given a practical optimization problem with multiple criteria, a common challenge is to provide an intuitive, fast and clear interface supporting the decision maker. But at the same time we want to avoid a high-handed, arbitrary simplification of the objectives. For convex objective functions we can achieve both by linearly approximating the efficient frontier in the objective space. This is used to enable the decision maker to ’navigate’ on the efficient solutions by interpolating the precalculated solutions in real time to satisfy the decision makers demands as good as possible. This ’navigation approach’ is established and proven to be successful. However, the approximation is unacceptably slow if the number of objectives is too large. However, in practical application we made an interesting observation: In cases of many objectives one usually needs just a small fraction of the approximation for doing the navigation and satisfying the decision maker’s demands. In the talk we present and explain this phenomenon and discuss how to use it to our benefit. For this we notice that the linear approximation is in fact a polyhedron with specific geometric properties. The analysis of it helps us to design an extension of the approximation allowing to apply it to larger number of objectives. The main ingredients for this are ideas from principal component analysis and singular value decomposition, respectively. Keywords: Efficient Frontier, Interactive Methods, Linear Approximation, Multi-Objective Optimization, Non-Linear Multi-Objective Optimization, Principal Component Analysis, Singular Value Decomposition. 307 FR2G (contributed) Proximity-Based Decision Rules for Multicriteria Model Predictive Control Problems Skulimowski, Andrzej M.J. (*) AGH UST The complexity of the problem how to define suitable decision rules for multicriteria decision-making problems occurring in model predictive control depends strongly on the nature of additional preference information available during the decision-making process. A common virtue of such problems is that the relation between the variability of the preference structure, the system parameter changes, and the computational capabilities of the problem solver play an important role. There also exist trade-offs between different solution approaches: an adequate control system model can yield an accurate solution but it might be never deployed because of the relatively quick changes in the system parameters or in the decision-maker’s preferences that exceed the computational and implementation capacities. Such problems occur e.g. in dynamic decision problems where the system parameters are derived from so-called big data streams or in discrete event control problems, where observable events are retrieved from the web, a monitoring, or a telecommunication system. Due to a large amount of data, the model of the system dynamics, the prediction, and decision rules must be simplified to an extent allowing their efficient processing and timely deployment of control. This paper will propose a methodology that refers to the anytime and continual computing paradigms and allows for a gradual simplification of the model by an appropriate discretization of time and system parameters to comply with the computational capabilities estimated ex-ante. Then we will show how to simplify the preference structures and define efficient proximity-based decision rules to cope with the limited decision-maker’s capability to calculate the solution in due time. This approach has been applied to recommend Pareto-optimal strategies of information society development described by a discrete-time linear non-stationary control system with parameters identified based on data from heterogeneous web sources. Keywords: Anytime Computing, Control, Multi-Criteria Model Predictive Control, Optimal Control, Preference Information, Preference Modeling, Proximity-Based Decision Rules, Reference Multifunction Method. 308 TH1A (plenary) Public Decisionmaking and Decision Conferencing Bana e Costa, Carlos (*) Technical University of Lisbon In this talk, I shall reflect on lessons learned from my professional involvement, as a decision-analyst and process consultant, in a variety of socio-technical processes, in which multicriteria value measurement techniques were used during decision conferences, with the participation of several types of social actors and stakeholders, with the common aim of facilitating better public decisionmaking. 309 WE1C (contributed) Quality of the Information Provided by Fuzzy Data in the Efficiency Evaluation of the Spanish Textile Industry Baeza-Sampere, Ismael (*) Universidad de Valencia Coll-Serrano, Vicente Universidad de Valencia Efficiency results obtained by applying conventional DEA models are usually used for Decision Making. These assessments are based on the assumption of certain data and, as noted by some authors in the academic literature, this assumption is not always acceptable. In these cases it is preferable to consider some kind of imprecision in the data. This is the aim of Fuzzy DEA models. The importance of the quality of the results is what leads us to propose a method to evaluate it. The comparison with the results of a crisp DEA model at the lower and at the upper values in the worst alpha-cut will allow us assess the quality of the results obtained when evaluating the efficiency of Spanish textile companies with a possibilistic model. Keywords: Data Envelopment Analysis, Fuzzy Mathematical Programming, Possibilistic Model. 310 MO3D (contributed) Quantifying Public Preferences for Evaluation Multifunctional Agriculture System, using Analytic Hierarchy Process Marqu´es, Inmaculada (*) P´erez-Salas, Jos´e Lu´ıs Segura Garc´ıa del R´ıo, Baldomero Universidad Polit´ecnica de Valencia Some agricultural systems in the urban fringe, maintain a fragile equilibrium between the consideration of agricultural space and/or periurban areas, subject to pressure by alternative use of the land (industrial, infrastructure, commercial and leisure centers, ..). Some of these agricultural systems are characterized as agricultural ecosystems that provide goods and services related to leisure and recreation, the process development beneficial to the environment, such as fixing CO2, the production of healthy and safe food, the preservation of natural and cultural heritage, ... The production of these goods and services derives mainly from practicing agriculture itself, and is generated simultaneously with the production of appropriate market goods. These are goods which do not have a market where farmers could sell them. The lack of de market can justifies the intervention of public administration. The difficult we are facing is the correct estimate of the social demand for these goods and services. An efficient policy should be attempt the public preference for non-market goods and services, and should be paid attention to the explicit measurement of public demand. Analytical Hierarchy Process can help assess the social priorities to integrate them into the decisions policies with the aimed to maximize the use of the agricultural system. In this paper we study social preferences for the functions of the land system of the Valencia urban fringe. We take representative residents of the study area and compare them with stakeholders. We must show, how far the preferences of the stakeholders are truly representative for the community as a whole. In both cases the priorities are calculated, and the aggregation of these is studied taking in to consideration the consistency of the individuals by adrresing their opinion of the preferences, on individual and aggregated level, studying the relation between them. Keywords: Agricultural Public Goods, Agriculture, Consistency, Decision Maker’s Preferences. 311 WE2B (contributed) Rank Reversal in the PROMETHEE Methods: a Summary of Recent Investigations De Smet, Yves (*) Universit´e libre de Bruxelles Multicriteria methods based on pairwise comparisons suffer from possible rank reversal occurrences when the set of alternatives is modified. Evidences of this distinctive feature have been found in AHP, ELECTRE and PROMETHEE. In this talk, we summarize the results of four related works done in the context of the PROMETHEE I and II methods (Mareschal, B. et al. (2008); Roland, J. et al. (2011); Verly, C. and De Smet, Y. (2013); Eppe, S. and De Smet, Y. (forthcoming)). First empirical tests are conducted on the basis of artificial data sets in order to quantify the likelihood of rank reversal instances. Then extensions to real data sets are presented. Conditions to avoid this outcome are studied. This leads us to develop theoretical bounds between flow scores differences that allow the detection of potential rank reversals. Then, extensions of the classic definition of rank reversal are considered; deletion of a non-discriminating criterion, addition of a copy of an alternative, etc. Finally, we investigate risks of manipulation on a simplified version of the PROMETHEE II ranking. Keywords: PROMETHEE, Rank Reversal. 312 TH2C (contributed) Ranking Candidates through Variable Convex Sequences of Weights Llamazares, Bonifacio (*) University of Valladolid Scoring rules are a well-known class of positional voting systems where fixed scores are assigned to the different ranks. Nevertheless, since the winners may change according to the scores used, the choice of the scoring vector is not obvious. For this reason, and by using the DEA methodology, several methods have been suggested so that each candidate may be evaluated with the most favorable scoring vector for him/her. In this work we propose a new positional voting system that allows to use different scoring vector for each candidate and avoid some shortcomings of the methods proposed in the literature. For this, we use constraints that assure convex sequences of weights. Keywords: Convex Sequences Of Weights, Data Envelopment Analysis, Positional Voting Systems, Scoring Rules. 313 TH2C (contributed) Re-Measuring the Happy Planet Index using DEA and MCDM Models Jablonsky, Josef (*) University of Economics Prague Happy Planet Index (HPI) is an aggregated index that measures the extent to which each nation produces long and happy lives per unit of environmental input. The HPI uses global data on life expectancy, experienced well-being, and ecological footprint in order to rank countries. The last HPI report was published in 2012 and it contains data for 151 countries of all continents. The aim of the paper is to re-calculate the HPI using DEA models and other multiple criteria decision making techniques and compare the given results. In addition the formulation of the problem is extended by some other criteria (GDP per capita, index of governance, etc.) and analyzes the impact of the new indicators on final ranking. MCDM methods evaluate alternatives (countries) according to the given set of criteria with respecting preferences of decision makers. Most of them allow ranking of alternatives according to aggregated indices defined by the methods (WSA, PROMETHEE II, TOPSIS, AHP, etc.). DEA models compare the countries with the best performers in the data set and measure the efficiency of transformation of multiple inputs into multiple outputs. Even they are based on different principles than MCDM methods they allow ranking of evaluated units according to their efficiency expressed by efficiency or super-efficiency scores. The paper analyzes both methodological approaches and compares their results. Numerical experiments are realized using original MS Excel add-ins for DEA models (DEA Excel solver) and MCDM methods (Sanna). Both applications can be downloaded from author’s web pages. Keywords: Data Envelopment Analysis, Multi-Criteria Decision Making, Software. 314 TU4F (contributed) Redundant Phenotype-Objective Space Mappings in Multi-Criterion Data-Clustering Handl, Julia (*) The University of Manchester John, Agah Niger Delta University Knowles, Joshua The University of Manchester Delattre and Hansen (1980) propose the definition of data clustering as a bicriterion optimization problem and describe an exact approach to bicriterion clustering, which uses branch-and-bound in combination with graph colouring to obtain the Pareto front for the minimization of maximum diameter and the maximization of minimum split. Due to limited computational resources at the time, the evaluation of the approach was limited to a handful of data sets of limited size. More recently, multiobjective evolutionary algorithms (MOEAs) have been used to provide more general implementations of multiobjective data clustering, which allow for the optimization of arbitrary clustering objectives (Handl and Knowles, 2007). Here, we are interested in a direct comparison of Delattre and Hansen’s method with an MOEA optimizing the same pair of objectives. Our comparison focuses on the objective values obtained as well as the external validity of the results obtained. An analysis in objective space shows that - as may be expected - the MOEA produces approximations to the Pareto front only, and that some of its solutions are dominated by those partitionings that are generated using the exact approach. On the other hand, the external validation of clustering solutions reveals a performance advantage for the MOEA, which seems counter-intuitive at first. Further investigation reveals that this finding arises from redundant phenotype-objective space mapping in certain formulations of data clustering. In particular, the concrete mathematical form of the minimum diameter and maximum split means that the search space of possible partitionings contains sets of different clustering solutions that project to the same points in objective space. The use of a linkage-based encoding in the MOEA (Handl and Knowles, 2007) implicitly deals with this redundancy by introducing an implicit bias towards high-density clusters. Keywords: Evolutionary Computations, Evolutionary Multi-Objective Optimization, Multi-Criteria Decision Analysis. 315 TU3E (contributed) Reference Point Methods and Approximation of Pareto Sets Goetzmann, Kai-Simon (*) TU Berlin Matuschke, Jannik TU Berlin B¨ using, Christina RWTH Aachen Stiller, Sebastian TU Berlin In theoretical computer science, research on multicriteria optimization problems usually revolves around the set of Pareto optimal solutions, or, for reasons of tractability, approximations of this set. In practical MCDM, on the other hand, methods that output a single solution are more widespread. Among these methods, reference point solutions and, as a special case, compromise programming, are especially popular. A reference point solution is the solution closest to a given reference point in the objective space. Compromise programming uses the ideal point as a reference point. We establish a link between reference point methods and the theory of approximate Pareto sets, by showing a polynomial equivalence of approximability: An approximate Pareto set can be efficiently constructed, if and only if there is an efficient algorithm that computes a reference point solution with a provable approximation guarantee. To illustrate our results, we lift some approximation techniques from single-criteria optimization to reference point methods, resulting in approximation algorithms for the Pareto sets of the multicriteria versions of some classical combinatorial optimization problems. Keywords: Achievement Scalarizing Functions, Combinatorial Optimization, Compromise Programming, Provable Approximation Guarantees, Reference Point. 316 TU2B (contributed) Relational Multi-Attribute Models in DEX Methodology Nejc, Trdin (*) Jozef Stefan Institute Bohanec, Marko Joˇzef Stefan Institute DEX is a qualitative decision support methodology aimed at evaluation and analysis of decision alternatives. The method is based on hierarchical decomposition of attributes, where leaf nodes of the hierarchy represent the input attributes of the alternative and the root attribute represents the final evaluation of a particular alternative. Furthermore, other intermediate attributes of the hierarchy represent logical combinations of lower attributes into higher level concepts. The aggregation in DEX models is governed by qualitative combinatorial tables, easily understandable as if-then rules. Many decision problems addressed in real life applications are based on some relational properties between at least two types of entities. For example, when evaluating a company, many departments of that company must be evaluated, too. The number of departments may vary in time, but each of the departments contributes to the evaluation of the whole company. Similar, but more complex example of relational properties is when assessing the reputational risk of a bank. Banks represent many counterparts, which serve many different products. These products are bought by several bank’s clients. The assessment of reputational risk thus depends on many relational connections. Currently, DEX only supports the evaluation of non-relational (“flat”) alternatives. In order to address relational problems, ad-hoc manual or programmatic manipulation of models is necessary, which adds additional burden to solving such problems. In this work we address the task of extending the multi-criteria decision making method DEX to relational decision problems. We will present a formal description of DEX and extend it to facilitate the treatment of relational decision alternatives. We will explain and justify the need for such an extension arising from practical applications and illustrate the use of the newly extended methodology on a real life example. Keywords: Preference Aggregation, Qualitative Multi-Criteria Model, Relational Modeling. 317 FR3F (contributed) Rewarding Decisions Beyond the Performance Targets Paradigm Sousa, Victor (*) Universidade do Porto Faustino, Ana Universidade do Porto A water resources management model was first approached using MCDM techniques in order to involve the representatives of the several uses of available water (in reservoirs or underground aquifers) in the design of rules to carry out that management. We were successful in having a model of optimization ruling management decisions, or “driving the system”, which was compatible with the stochastic process of inflows to the system and used, in a particular way, dynamic programming and multicriteria analysis. This model worked under the establishment of independent performance targets and penalty functions for each selected use of water. The optimization DP through a given elapsed time was made according to previous accepted rules settling those targets through all different seasons and specific conditions. The optimal solution provided a driving policy of the reservoir by minimizing the global penalties for the accumulated shortages, as agreed with the representatives of the different users of water. This “performance targets paradigm” has been questioned by a few water resources specialists and an alternative was presented by Abbas and Matheson, 2009, mainly by introducing and developing the concept of “value aspiration equivalent” to overcome the artificial independence of targets and regulate trade-offs among the performance objectives. Our first adjustment of our model inspired by this new paradigm led to the consideration of “reward functions” along the “penalty functions”, those set of functions unveiling a good architecture to the trading-off system. Results were very similar to the first model under shortage situations, but provided non negligible advantages under sufficiency of resources. Our most recent work proposes a methodology to merge those two types of functions along with optimisation techniques, obtaining interesting results either under shortage or abundance of resources. Keywords: Penalty And Reward Functions, Performance Targets, Water Resources. 318 FR3F (contributed) Risk Assessment of Large Scale Hydroelectric Plant Projects using Fuzzy Analytical Hierarchy Process Ribas, Jose Roberto (*) Universidade Federal do Rio de Janeiro Severo, Juliana Ribas Delta Energia Sohler, Flavio Furnas Centrais Eletricas S/A The difference between planned and actual outcomes of a project is associated to risk events, in order to avoid poor performance; the construction industry is required to manage such risks. Research and practice has shown that construction of power plants embeds many risks. This study presents a method that integrates the expert subjective opinion and a multicriteria method, aiming at setting up hierarchical risk event structures related to the most relevant service packs of the project. Such information was captured by means of in-depth interviews and content analysis. The strength of one event risk against the other was assessed through pairwise comparisons using a nine-point Saaty’s scale. The adjustment for low precision of subjective opinion and the quantitative assessment of the event risks weights were achieved by means of a fuzzy analytical hierarchy process (FAHP). The risks associated to Santo Antonio hydroelectric plant were assessed. This plant, which is located in the Amazon rainforest region, has 3,580 MW of installed capacity, and represents one of the main challenges to the Brazilian engineering. Mainly because it will work with bulb turbine technology and will generate energy from the flow of the river Madeira itself, without the need of creating a large reservoir. This complex large scale project has five major event risk-related to five service packs, as specified by the experts. A comparison was made between the risk events weights estimated through the FAHP technique for two segments of participants: the construction managers and the owner’s engineers. It was concluded that some of the differences were due to the opposing risk profiles of the two groups, as a consequence of their obligations as established by a turnkey lump sum contract. Keywords: FAHP, Hydroelectric Plant, Risk Analysis. 319 FR3D (invited) Risk-Averse Bi-Objective Optimization via Stochastic Dominance under Transaction Costs Gutjahr, Walter (*) University of Vienna A common approach to bi-objective stochastic optimization determines the set of “expected value efficient solutions” (Pareto optima w.r.t. the expectations of the objectives). However, this consideration presupposes a risk-neutral decision maker. In the presence of risk aversion, an analysis becomes much more complex. That is why authors as Caballero et al. (2004) questioned the suitability of the “multi-objective method” in this context at all. The talk proposes an alternative approach based on the concept of multivariate stochastic dominance constraints, as it has recently been elaborated to address risk aversion w.r.t. more than one outcome variable. This concept uses a benchmark solution and optimizes an objective function over all solutions stochastically dominating the benchmark. Whereas the most common multivariate stochastic order in the stochastic programming literature, the positive linear stochastic order, implicitly assumes a perfect market allowing to trade objective outcomes against each other, the present talk introduces a class of stochastic dominance relations based on the more realistic assumption that trading is only possible under transaction costs. The introduced class of dominance relations can be shown to contain both the positive linear order and the so-called component order as boundary cases. The variant for linear transaction costs is then applied to the special case of separable, piecewise linear, nondecreasing and concave utility functions. Mathematical programming formulations are derived. Finally, the application of the approach is outlined for a bi-objective stochastic covering location problem. For the computational solution, exact methods reyling on branch-and-bound as well as multiobjective evolutionary metaheuristics as NSGA-II are discussed. For a small real-world disaster management instance analyzed in a previous publication under risk neutrality, the exact Pareto frontier is determined in the risk-averse framework. Keywords: Covering Location Problem, Multi-Objective Optimization, Stochastic Dominance Constraints, Stochastic Programming, Transaction Costs. 320 FR2C (invited) Robust Elicitation of a Qualitative Ranking Model using Inconsistent Data Ouerdane, Wassila (*) Ecole Central Paris Liu, Jinyan Ecole Centrale Paris Mousseau, Vincent Ecole Centrale Paris Ranking with Multiple reference Points, namely RMP method is a fam- ily of qualitative aggregation models for multi-criteria decision aid (MCDA). In this work, we are more concerned about a Simplified RMP model (terms to S-RMP), which is an outranking method towards ranking prob- lem in MCDA. The aggregation proceeds by decomposing additively the importance relation on the set of criteria and it takes into account a lexicographic order of dictator reference points. A S-RMP model can be elicited from pairwise comparisons provided by the Decision Maker (DM). However, the current S-RMP elicitation algorithms are not designed to consider specifically the situations where the information provided by the DM contains inconsistencies. To tackle this problem, we propose a new version of the elicitation algo- rithm model by processing the inconsistency of pairwise comparison state- ments under the prudence principle. This new elicitation algorithm is for- mulated by a mixed integer linear problem. To test and evaluate the new algorithm’s performance, we implement the different elicitation algorithms for SRMP model in Java and solve the problem with CPLEX Optimizer. In addition, the programs have been structured into several modules to achieve various functions and the Input/Output is compatible with the XMCDA data standard defined via a XML schema. We design a series of numerical ex- periments for the new algorithm. The empirical results show that the new elicitation algorithm for S-RMP model has a greater ability to deal with the inconsistencies. For future research, we plan to extend this work to the case in which multiple decision makers provide pairwise comparisons; indeed it is similar to the inconsistency problem in that the multiple DMs may provide conflicting preference statements. Keywords: Inconsistency, Multi-Criteria Decision Aid, Pairwise Comparison, Reference Point, Robust Elicitation. 321 FR2C (invited) Robust Ordinal Regression in Geographical Information Systems Greco, Salvatore (*) Corrente, Salvatore University of Catania Boggia, Antonio Massei, Gianluca University of Perugia Slowinski, Roman Poznan University of Technology Multicriteria Spatial Decision Support Systems (MC-SDSS) couple Geographic Information Systems (GIS) and Multicriteria Decision Aiding (MCDA). On the part of MCDA, we propose to use Robust Ordinal Regression (ROR) methodology. ROR considers the whole set of utility functions that are compatible with the preference information provided by the Decision Maker (DM). With this aim, ROR presents the necessary preference, which holds when an alternative a is at least as good as another alternative b for all compatible utility functions, and the possible preference, which holds when an alternative a is at least as good as another alternative b for at least one compatible utility function. ROR has been adapted to deal with ranking and sorting problems, and it can also be used when preferences are represented by an outranking relation. Using ROR one can also analyze the possible rank of the considered alternative through the so-called Extreme Ranking Analysis. ROR permits also to consider interactions among criteria using either the Choquet integral or an “enriched” additive utility function including positive and negative synergy components. We discuss the benefits of application of ROR to MC-SDSS, and we illustrate this application by a case study on land use. Keywords: GIS, Multi-Criteria Decision Aid, Robust Ordinal Regression. 322 TU2F (contributed) Robust Selection of Environmental Quality Metrics based on a Multi-Criteria Pareto Data Mining Process Udias, Angel (*) Joint Research Centre Galbiati, Lorenzo Ag`encia Catalana de l’Aigua Cano, Javier Chiazzaro, Mauricio Rey Juan Carlos University Redchuk, Andr´es Universidad Aut´onoma de Chile Sustainability of water resources has become a challenging problem worldwide. In this context, the development and application of new political and management strategies, aimed at overcoming the growing problem of scarcity of fresh water, is of vital importance. Specifically, planning and management of water resources requires making decisions characterized by multiple objectives and stakeholder groups, according to various criteria measuring water quality. This would typically imply deciding what action should be accomplished at each of a series of spatial locations along the river, as well as their effects on such criteria. Hence, these problems, called Spatial-MCDM methods, extend traditional multicriteria decision methods by including explicit geographic information about the alternatives and quality objectives. However, Spatial-MCDM methods are not universally accepted as a standard way to assess the quality of a river, since, for a specific indicator, different techniques for spatial data aggregation are conceivable, possibly yielding different tradeoffs and decision strategies. Throughout an aggregation process, the original spatial data are reduced to a small number of data elements. Moreover, the aggregation process may alter the statistical and spatial characteristics of the data. To study these effects, we apply different aggregation techniques on a catchment scenario within a hydrological MCDM methodology. Such procedure is aimed at finding an Program of Measures (PoMs), fulfilling the EU Water Framework Directive objectives, and consisting of: 1) A water quality model (to simulate the effects of the PoMs); 2) A Multi-Objective Evolutionary Algorithm, used to identify efficient trade-offs between PoMs’ costs and water quality; and 3) A data mining process of the Pareto set, in order to extract knowledge from Pareto-optimal decisions. A comparison using different aggregation techniques is presented, showing which metrics yield more robust results. Keywords: Data Mining, Environmental Decision Making, Evolutionary Multi-Objective Optimization, Multi-Objective Optimization, Pareto Optimal Solutions, Spatial MultiCriteria Evaluation, Water Resources Management. 323 WE1G (contributed) Robust Weighted Solutions for Multi-Objective Linear Programming Ogryczak, Wlodzimierz (*) Warsaw University of Technology In Multi-Objective Linear Programming (MOLP) any efficient solution can be found by the weighting approach with some positive weights allocated to several objectives. The weights settings represent preferences model thus involving impreciseness and uncertainties. The resulting weighted average performance may be lower than expected. Several approaches have been developed to deal with uncertain or imprecise data. We analyze robust approaches to the weighted averages of objectives where the weights are known imprecisely. For the case of unlimited weights selection the worst case weighted average becomes the worst outcome (max-min solution). In the case of box impreciseness, the worst case weighted average is a parametric generalization of the worst outcome to the so-called tail average. We show that while considering robust models allowing the weights to vary only within given intervals, the tail average represents the robust solution for only upper bounded weights. For any arbitrary intervals of weights the corresponding robust solution may be expressed by the optimization of appropriately combined weighted average and tail average of objectives thus remaining easily implementable with auxiliary linear inequalities. Keywords: Linear Programming, Multi-Objective Optimization, Robustness And Sensitivity Analysis. 324 MO5H (poster) Routing System Optimization – Case Study Fuentes Rojas, Ever Angel (*) Universidad Libre Mart´ınez Gamboa, Jeyson Andres Universidad Libre The requirements demanded by the society make companies optimize its distribution processes as an improvement strategy in order to be competitive and survive in the market for satisfying customers’ needs. Plastics industry is no stranger to this situation that is why it is necessary to optimize the routing system of a company type, which has its activity center in Puente Aranda, Bogota- Colombia, and has decided to reserve its business name for this research. It has as starting point the results that were got in the diagnosis test. Thereafter data were analyzed through the use of administrative tools as: Cause- effect diagram, Pareto, ABC classification and SWOT matrix. These ones allowed establishing the most important variables: Costs, distance and response time. This form is achieved to optimize the routing system by the application of the mathematical model of the travel salesman problem (TSP) whose solution was found through the use of programs as Gusek and Grafos. These programs offer an optimal solution with costs reduced 32% and service level increased 26%. Prospective system is implementing through the use of 3 scenarios which were tested with simulation techniques in order to select a decision. Finally financial indicators serve as support of the decisions made by the company representing the best course of action to follow with lower costs and good percentages in the internal rate of return (IRR) of the investments. Keywords: Costs, Graph Theory, Optimization, Plastic, Travelling Salesman. 325 FR2F (invited) Satisfying Multiple Objectives in Wastewater Treatment Network Design Using a Diversity Preserving Evolutionary Algorithm Chmielewski, Hana (*) North Carolina State University Ranjithan, S. Ranji North Carolina State University There is a growing viewpoint in multi-objective civil infrastructure modeling that the success of an optimization method depends on its ability to handle unmodeled as well as modeled objectives. Solutions that are optimal in one or more objectives may not be feasible, or reflect unmodeled priorities. Using a diversity-preserving evolutionary algorithm (EA) is a step towards increasing algorithm robustness in producing ?nal solution sets that are more likely to contain diverse, high-quality alternatives that address the modeled and unmodeled issues facing decisionmakers. A regional wastewater network case study in DuPage County, Illinois, can be modeled to exhibit the cost, environmental, and resilience tradeoffs experienced in wastewater management. The case study area includes wastewater sources, potential treatment plant locations, and possible interceptor routes between sites. In addition to minimizing costs, a newly-defined network decentralization objective is added to carry out a two-objective optimization. The diversity of the decision space is explored using two approaches: a new diversity preserving EA approach called Diversity Ranking Evolutionary Multi-objective Algorithm (DREMA), and a mixed-integer linear programming (MILP) approach, both designed to find solutions that are maximally different in the decision space. To simulate decision-making scenarios in which new priorities may arise mid-analysis, the performance of the two-objective results of DREMA and MILP are compared against several unmodeled issues, including a third network resilience objective. Compared to MILP, DREMA produces solutions competitive in cost, and competitive or superior in the decentralization objective and resilience performance. Both methods produce highly diverse solution sets, as measured by two diversity metrics, and DREMA does so without the inconvenience of objective function reformulation after each alternative. Keywords: Decision Space Diversity, Evolutionary Computations, Multi-Objective Optimization, Network Resilience, Water Resources Management. 326 TU5D (invited) Scenario-based Multi-Criteria Decision Support for Robust Humanitarian Relief Supply Chains Comes, Tina (*) University of Agder Sch¨atter, Frank Karlsruhe Institute of Technology Humanitarian relief logistics are among the most challenging problems in logistics and supply chain management. We present an approach combining decision-analytical techniques, optimization and scenario-based reasoning and demonstrate our results by a case-study (Haiti earthquake 2010). Our approach focuses on a facility location problem (FLP), a strategic decision problem, which is essential for designing supply chain networks . Risks in an environment hit by a disaster are hard to assess. The environment is characterized by high complexity and uncertainties, and information is sparse or even lacking particularly in the early phases of the disaster. As typically, various organizations need to collaborate, multiple objectives and preferences need to be taken into account to facilitate consensus building and coordination. In these situations, techniques from Multi-Attribute DecisionMaking (MADM) have proven useful. These assume that a small set of alternatives has been identified. Typically, these alternatives are created intuitively, without relying on computational support, as decision makers need to respond quickly under stress and time pressure. We present a systematic way to generate and evaluate alternatives to support the design of robust relief supply chains by combining optimization and MADM. Scenarios are used to explore different developments of the disaster situation in a dynamic manner. Fundamental uncertainties after the Haiti earthquake comprise the disaster’s impact on the population, infrastructure , available resources, or damages and disruptions. These most prominent uncertainties are captured in a set of scenarios and are evaluated by measures of MADM. The use of a scenario-based MADM combined with techniques from operations research to solve the FLP enables identifying robust alternatives (in terms of stability and quality) that perform well under various scenarios and thus enable coping with fundamentally different situation developments. Keywords: Facility Location Problem, Humanitarian Logistics, Multi-Criteria Decision Aid, Robust Supply Chains, Scenarios. 327 TU5G (invited) Searching Graphs with Lexicographic Goal References Mandow, Lorenzo (*) Universidad de M´alaga P´erez-de-la-Cruz, Jos´e L. Universidad de M´alaga Pulido, Francisco Javier Universidad de Malaga We consider the problem of searching for paths in a multiobjective graph from given source and destination nodes. More precisely, we search for the subset of all Paretooptimal paths that satisfy a set of lexicographic goals, i.e. goals grouped in preemptive priority levels. We assume additive costs, and minimize the weighted deviation of goals for each priority level. We propose LEXGO*, a new algorithm for the efficient solution of this problem. LEXGO* is a label-setting algorithm that generates partial paths from the start node to the destination. Each partial path is in turn expanded, filtered, or pruned. Conditions for filtering and prunning are defined in such a way that discarded paths are among those that would never generate an optimal solution. Hence, it can be proved that LEXGO* is admissible (i. e., it finds exactly the set of best solutions). The algorithm can exploit heuristic information to improve search efficiency. It can be also proved that, under certain conditions, the number of explored labels can be reduced using better heuristics. Some experiments have been carried out to analyze the performance of LEXGO*. Random problem have been generated on a grid and solved (i) by LEXGO*; (ii) by the optimal multiobjective algorithm NAMOA*, first generating all Pareto-optimal solutions and then selecting among them those that satisfy the goals. Experiments show that: (a) when few solutions exist, LEXGO* achieves a substantial reduction of execution time over procedure (ii); (b) when many solutions exist, execution times are very similar. Keywords: Artificial Intelligence, Goal Programming, Multi-Objective Optimization, Pareto Optimal Solutions. 328 MO4G (contributed) Sensitivity and Efficiency of Multiple Criteria Alternatives of the Model ’Revenue - Risk’ Sarkisyan, Rafael (*) Moscow University of Engineering Eskova, Olga Moscow University of Engineering Employing a multiple criteria logic for the analysis of investment solutions on the basis of the model “Revenue - Risk” allows to achieve an acceptable compromise between the competing values of revenue and risk criteria with regard for the key economic factors of a situation and its development trends. The strategy of a multiple criteria optimization is based on the concept of preferences and utility, a consecutive revelation and description of which in the framework of the friendly interactive procedure serves an effective means for reducing the uncertainty and non-comparableness of alternatives. With regard to the model “Revenue - Risk” it allows to efficiently combine scales for measuring the revenue and risk into a significant scale for measuring the degree of their domination. The transformation of the initial space of estimates R0 = [0, ρp ] × [0, rp ], where ρp and rp are potentially achievable values of the risk ρ and the revenue r respectively, into the space F0 = {f ∈ E+2 /f = (f1 , f2 )T , f1 = ρp − ρ, f2 = r} with the coordinated directions of preferences increase, allows to bring in as the main tools of analysis and optimization the measures of sensitivity σ(f ) = f T ∇u(f )/u(f ) and of efficiency (or return) ν(f ) = f T ∇u(f ) = σ(f )u(f ) of the concave increasing on F0 differentiable utility function u(f ). Investigation of the behavior of the functions ν(f ) and σ(f ) on the indifference curve u(f ) = const, generates the perspective estimates trajectory f c (λ) with the maximum values of ν(f c ) and σ(f c ) and with the level of the coordinates interaction function ϕ(f c ) = −f cT Hf c = (1 − λ)ν(f c ); λ ≤ 1 is the Lagrange multiplier, H is the Hessian matrix of the function u(f ). A construction and interpretation of these correlations on specific numerical data are illustrated on the example of the quadratic approximation u(f ) = aT f + (1/2)f T Hf . Keywords: Measures Of Sensitivity And Efficiency, Multi-Criteria Logic, Preferences And Utilities, Revenue-Risk Model. 329 WE1D (contributed) Simulation-Based Innovization for Improving Productivity and Energy Efficiency of Manufacturing Systems Ng, Amos (*) Goienetxea, Ainhoa Urenda Moris, Matias University of Sk¨ovde Deb, Kalyanmoy Michigan State University Dudas, Catarina Volvo Group Trucks Technology This paper introduces an extended simulation-based innovization (SBI) methodology which combines the use of discrete-event simulation with evolutionary multi-objective optimization (EMO), multi-criteria decision making, data mining and advanced data visualization for eliciting knowledge for intelligent optimization and informed decision making in systems engineering problems. Special focus will be paid on how the methodology can be applied to simultaneously improve the productivity and energy efficiency of manufacturing systems producing discrete products, illustrated through several realworld case studies conducted in automotive industry. The SBI methodology is based on the innovization procedure, originally introduced for unveiling new and innovative design principles in engineering design problems. Although the innovization method is based on multi-objective optimization and post-optimality analyses of optimized solutions, it stretches the scope beyond an optimization task and attempts to discover new design/operational rules/principles relating to decision variables and objectives, so that a deeper understanding of the problem can be obtained. The uniqueness of the extended SBI approach presented in this talk lies on the applications of MCDM techniques, particularly the reference-point based approach, so that the decision maker can control both the optimization task and the knowledge elicitation process towards his/her preferred regions in the objective space. As illustrated with the experience learnt from the decision making process in the industrial case studies, the paper will convince that SBI not only helping production managers/engineers to explore optimal design and decision variable settings, but also gaining better knowledge and insight about production systems engineering in general. Keywords: Data Mining, Decision Maker’s Preferences, Energy Efficiency, Evolutionary Multi-Objective Optimization, Production, Sustainability. 330 MO2C (contributed) Solving a Mutiobjective Dynamic Problem with a Stochastic Transition Function and Stochastic Objective Function by Interactive Procedure Nowak, Maciej (*) University of Economics in Katowice Trzaskalik, Tadeusz University of Economics in Katowice Real-world decision problems are often dynamic. In such case, the decision-maker makes a sequence of choices in order to maximize the total benefit resulting from the benefits obtained at the subsequent stages of the process. However, the longer is the planning horizon, the less certain is the result. Furthermore, serious decision problems involve multiple and conflicting criteria. In the paper a multiobjective dynamic decision-making problem is considered. We assume that both the transition function as well as objective functions are stochastic. The decision maker goal is to find the most preferred solution. The procedure we propose uses Bellman principle for finding the efficient set, and then interactive approach for identifying the final solution. A numerical example and a realworld application are presented to illustrate the applicability of the proposed technique. Keywords: Dynamic Programming, Interactive Methods, Multi-Objective Optimization, Stochastic Programming. 331 TH2D (invited) Solving Fuzzy Mathematical Programming Problems by a Parametric Approach: a Historical View Cruz Corona, Carlos (*) University of Granada Coelho Silva, Ricardo Federal University of Sao Paulo Verdegay Galdeano, Jos´e Luis University of Granada In 1974 Tanaka defined the concept of fuzzy mathematical programming applying the theory of fuzzy sets to decision problems based on the concept of Bellman and Zadeh for decision making under fuzzy conditions. Also, in those years Negoita and Sularia shown that the problem of determination of a maximizing decision can be reduced to a mathematical programming problem. So then, a lot of models, solutions and real-world applications were developed. Using a dual parametric approach was one of the most interesting idea that appeared in those days for solving mathematical programming problems. In this work, we present a historical view of this approach, and describe some of their applications to different problems. Keywords: Fuzzy Mathematical Programming. 332 TU5F (contributed) Solving Linear Fractional Programming Problems Using Multiobjective Optimization Techniques Hasannasab , Maryam (*) Kharazmi University of Tehran Ehrgott, Matthias Lancaster University This paper describes an approach for solving linear fractional programming problems via reformulation as a bi-objective linear programming (BOLP) problem. We apply Benson’s outer approximation algorithm for BOLP as presented in Ehrgott, Loehne, and Shao (2011) in order to solve the BOFP problem and derive the optimal solution for the LFP in the process. We contrast the performance of the primal and dual formulations of Benson’s algorithm with established algorithms for the LFP on various test problems drawn from the literature. Keywords: Fractional Programming, Multi-Objective Optimization, Outer Approximation. 333 WE2B (contributed) Solving Ranking Problems with ELECTRE-III in Case of Hierarchical Family of Criteria Del Vasto, Luis (*) Universitat Rovira i Virgili Slowinski, Roman Poznan University of Technology Valls, Aida Universitat Rovira i Virgili Zielniewicz, Piotr Poznan University of Technology The current ELECTRE-III method deals with criteria defined on a common level only. In some real-world problems, however, criteria are organized hierarchically. The leaves of the hierarchy tree correspond to elementary criteria on which a finite set of alternatives is directly evaluated. Then, elementary criteria are grouped in more general (intermediate) criteria, up to a root node that represents a final decision. We propose an extension of the ELECTRE-III method to deal with the hierarchical organization of criteria, called ELECTRE-III-H. At a node being the parent node for a given subset of elementary criteria, the two step process typical for ELECTRE methods is applied: (1) construction of a credibility matrix based on the marginal concordance and discordance indices, and (2) exploitation of the credibility matrix via distillation, obtaining a partial preorder of alternatives. This partial preorder is interpreted as a result of evaluation of the set of alternatives by the intermediate criterion corresponding to the parent node. Then, in order to aggregate the set of intermediate criteria with a common parent node of the higher level, one must take into account evaluations of alternatives in form of partial preorders. To aggregate partial orders using the concordance and non-discordance tests of ELECTRE-III, we first adapt the marginal concordance and discordance indices to handle all binary preference relations (indifference, incomparability and preference) from a given partial order, and then we use these marginal indices to calculate the credibility degree. The distillation of the credibility matrix does not change, and it results in a partial preorder at the parent node of the higher level. The aggregation proceeds until the root of the tree, where the final partial order is obtained. A case study on ranking Websites from different Touristic City Brands is presented. This work is supported by Spanish projects DAMASK (TIN2009-11005) and SHADE (TIN2012-34369). Keywords: Hierarchical Criteria, Multi-Criteria Decision Making, Outranking Methods, Partial Preorder. 334 WE2F (invited) Solving Robust Bicriteria Shortest Path Problems Raith, Andrea (*) The University of Auckland Kuhn, Kenneth RAND Corporation Schmidt, Marie 0.2cm Sch¨obel, Anita Georg-August University G¨ottingen Finding robust solutions of optimization problems is important in practice and well researched for single-objective problems. However, in many applications optimization problems have two or more objective functions. Here, we extend robustness concepts to a bicriteria optimization problem with one uncertain objective. This work is motivated by the application of routing airplanes through convective weather, where two conflicting objective functions are given by the length and the risk of the chosen route. One can identify efficient aircraft routes as solutions of a bicriteria shortest path problem. Pilots can select their preferred route among those efficient routes depending on how risk-averse they are. While the length objective is known, the risk associated with any route segment depends on weather forecasts and is therefore uncertain. We extend standard robustness concepts to the bicriteria case with one uncertain objective and we also propose new robustness concepts. The robustness concepts are analyzed for several types of uncertainty sets and the relationship between different concepts is identified. We introduce methods to solve robust bicriteria shortest path problems for our different robustness concepts and show computational results. Keywords: Robustness And Sensitivity Analysis, Shortest Path. 335 MO2E (contributed) Some Epistemological Considerations of the Goal Programming Model Aouni, Belaid (*) Laurentian University The Goal Programming (GP) is the most known model of the multi-objectives programming tools. This model is continually fed with theoretical developments and new applications with resounding success. It is a distance function model that aggregates simultaneously several conflicting and incommensurable dimensions where the deviations between the achievement and the aspiration levels are to be minimized. Considering several dimensions enhances the legitimacy of studying the decision-making problem from different aspects and facets. In fact, the GP model is based on a satisfying philosophy where the decision-maker rationality is bounded and the obtained recommendation (solution) is the one of the best compromise that fits with his/her value system and preferences. Despite that the GP model is not a pure axiomatic model but it does respect the general principals of mathematical modeling. The aim of this paper is to look at the GP model epistemological dimensions and to highlight its contribution to the multi-criteria decision aid paradigm. Keywords: Epistemology, Goal Programming, Multi-Criteria Decision Aid, Preference Modeling, Satisfying Philosophy. 336 TH3G (invited) Some Remarks about Multi-Criteria Methods into Planning for Use and Services of Natural Areas Ant´on, Jos´e (*) Andina, Diego Grau, Juan S´anchez, Maria Tarquis, Ana Universidad Polit´ecnica de Madrid Cisneros, Jos´e Manuel Universidad Nacional de R´ıo Cuarto Portait, Roland ´ ESSEC Grande Ecole de Commerce The authors have adapted with others some simple models of multi-criteria methods for planning about general uses of lands for some areas, or about special actions for them in Argentine, or formerly about facilities for transport or soil qualification in Spain, and they tell here some general comments. A planning field must have a technical doctrine to know its reality, evolution, actions and effects, and convenient multi-criteria modeling may aid to improve realistic appreciations. For soils of Madrid province they added to a FAO style worse criteria qualification a weighted functional evaluation and a necessary intermediate fuzzy assessment. They had used previously a related weighted evaluation for Spanish banks at 1995/96 transition years. For transport, that uses internal rate of return formats incorporating bayesian elicitations for future, authors had compromise programming experience about new R3 highway, where crisis causes at present that tolls are deterring users, and for the high speed train line Madrid-Valencia, in which with ELECTRE-I, and also with AHP a bit differently, long term effects were added to the official twenty five year horizon standard transport planning evaluations, and in fact a good real solution was build and put in service in that sense. For election of solutions for garbage in rural areas in Salta high valleys in Argentina they added a PROMETHEE-I method with ELECTRE inspired weights for the various criteria that appear involved. Such methods with weights elicited for long term planning, tried in various cases, matched for main uses of soils in Chaco Salte˜ no Argentine region, but for Arroyos Menores loess area South of Rio Cuarto in C´ordoba province the erosions required to focus more on hydraulic management and soil conservation, by incorporating them in use of soil alternatives. Towards shorter term horizon they experimented with continuous multi-criteria methods for more detailed planning in a sub area near Rio Cuarto. Keywords: Analytic Hierarchy Process, Compromise Programming, Multi-Criteria Decision Making, Outranking Methods. 337 WE1B (contributed) Standardization of Classroom Furniture in a Higher Education Institution Salomon, Valerio (*) Sao Paulo State University Alonso, Pedro University of Seville Marins, Fernando Sao Paulo State University As any other organization, higher education institutions have to choose resources to perform their process. Furniture used in laboratories, offices, and even in the classrooms are an important kind of resource. When there are two or more feasible alternatives for furniture, choosing the best furniture becomes a decision problem. One can make a mono-criterion decision, for instance, basing only in costs. Better than that, furniture selection can be modeled as a multi-criteria decision problem, incorporating attributes as user’s comfort and space usage. This work presents how a Brazilian higher education institution applied a multi-criteria decision-making method to standardize furniture to its classrooms. The use of four criteria, being two quantitative, enhancing the learning of multi-criteria decision-making is a major contribution of this paper. Keywords: Analytic Hierarchy Process, Furniture, Higher Education Institution. 338 FR3D (contributed) Stochastic Approch Versus Optimizition Over Efficient Set Mebrek, Fatma (*) ENSTP Chaabane, Djamal USTHB In this paper we study the problem of optimizing a linear function over an integer efficient solution set of a Multiple objective Stochastic Integer Linear Programming problem (MOSILP). Once the problem is converted into a deterministic one by adapting the 2levels recourse approach, a new pivoting technique is applied to generate an optimal efficient solution without having to enumerate all of them. This method combines both techniques, the L-Shaped method and the the combined method developed in [1]. Two didactic examples are given to illustrate different steps of our algorithm as well as mentioning the missed efficient solution when applying the algorithm in [2]. Keywords: Multi-Objective Optimization, Stochastic Linear Program. 339 TH3B (invited) Stochastic Ordinal Regression for Multiple Criteria Sorting Problems Kadzinski, Milosz (*) Pozna´ n University of Technology Tervonen, Tommi Erasmus University Rotterdam We present a new approach for multiple criteria sorting problems. We consider sorting procedures applying general additive value functions compatible with the given assignment examples. For the decision alternatives, we provide four types of results: (1) necessary and possible assignments from Robust Ordinal Regression (ROR), (2) class acceptability indices from a suitably adapted Stochastic Multicriteria Acceptability Analysis (SMAA) model, (3) necessary and possible assignment-based preference relations, and (4) assignment-based pair-wise outranking indices. In this way, we combine ordinal and stochastic analysis a unified decision aiding framework. ROR methods provide recommendations such as “depending on the chosen compatible instance of the preference model, the alternative is assigned to class medium or good or best” and “irrespective of the compatible model instance, the alternative is assigned to class bad”. However, our experiences indicate that the possible assignment is often rather wide, whereas the necessary assignment is often empty. In this perspective, it is useful to answer the question about how possible it is for a particular alternative to be assigned to each class. On the other hand, although the SMAA class acceptability indices (CAIs) can tractably be estimated to a reasonable accuracy, they are not exact. Thus, it is desirable to analyze CAIs in the context of the outcomes of ROR to provide information on which assignments occur with all, some or no compatible preference models. Furthermore, the sole SMAA method for multiple criteria sorting problems extends ELECTRE TRI. It is a pseudo-criterion based method, which requires preference information in the form of density functions for the class profiles, cutting level, and weights. Their elicitation may be too demanding for a DM. To address this problem, we adapt SMAA-TRI to apply general monotone value functions as the preference model. Keywords: Multi-Attribute Utility Theory, Robust Ordinal Regression, Robustness And Sensitivity Analysis, SMAA, UTADIS. 340 TU2C (contributed) Strategic Multi-Criteria Decision Analysis Montibeller, Gilberto (*) London School of Economics The choice of successful strategic options is at the core of organizational success. There are many challenges confronted by decision makers in these contexts. Firstly, they typically involve high levels of uncertainty, which may have a huge impact on the potential performances of the strategic options being considered. Secondly, usually there are many powerful stakeholder involved, either playing with the internal politics of private companies, or powerful interest groups inside and outside public organization. Thirdly, there is an increasing trend for participation in decisions, either because it increases the commitment to the solution chosen, or because it provides procedural justice and thus legitimacy to the decision. Fourthly, the planning horizon for many of those strategic decisions is typically very long. Fifthly, any strategic decision typically involves the pursuing of multiple and conflicting objectives, and the setting up of strategic priorities. Finally, each strategic option is a complex package of actions and sub-policies, so evaluating their performance is far from trivial. The answer to these challenges is, in my view, to adapt our methods and enhance our tools, borrowing ideas from different fields that have dealt with some of the issues involved in strategic decision making. Recent conceptual advances in Decision Sciences provide MCDA researchers and practitioners with more powerful toolkits than ever before. In this talk I present some of these trends and suggest opportunities for research and practice for Strategic Multi-Criteria Decision Analysis. Keywords: Decision Analysis, Problem Structuring, Robustness And Sensitivity Analysis, Strategic Decisions, Strategic Planning. 341 FR2B (contributed) Structural Modeling Approach to Multiple Criteria Problems with Interrelated Components Michnik, Jerzy (*) University of Economics in Katowice The majority of MCDM methods follows the assumption of independence between criteria. Yet, in many cases the interrelations between criteria cannot be neglected. There are the number of methods that have been developed in the field of structural modeling that can cope with interrelations between the components of analyzed system and enrich the modeling approaches of MCDM. This work presents the comparison of the four such methods: Analytic Network Process (ANP), Decision Making Trial and Evaluation Laboratory (DEMATEL), Composite Importance (CI) and Weighted Influence Non-linear Gauge System (WINGS). Keywords: Analytic Network Process, Composite Importance, DEMATEL, Interrelations, Multi-Criteria Decision Analysis, Structural Modeling, WINGS. 342 TH3F (contributed) Structured Development of Automotive Electric/Electronic Architectures Using Evolutionary Optimization Dietermann, Ansgar (*) Dresden University of Technology B¨aker, Bernard Dresden University of Technology Automotive electric/electronic architectures describe all electric and electrical components in a vehicle as well as their interconnection and reciprocal effects. The complexity of these systems is increasing continuously as new electric/electronic functionalities are added to the vehicle, and existing mechanical functionalities are exchanged by electrical ones. The electrical power train represents both of these challenges, as the combustion engine is replaced by an electric engine and new functionalities like the high voltage level are introduced. The rising complexity influences the development methodology of the architectures, which is still characterized to a huge degree by the expert knowledge and subjectivity of the specific developer. The first step in developing a new e/e-architecture is setting up the architectural concept. In doing this, the basic and most important decisions are made, which define the quality of the finished architecture to a huge extent. A developing method has to be found that structures the e/e-architecture development generically, thus increasing the transparency and traceability of the made decisions, and that leads to an optimized architecture concept to ensure the high quality of the later e/e-architecture. A multiple criteria decision making method based on evolutionary optimization of the logical allocation and functional integration of technical modules to e/e-architecture components is presented. The used example is the component set and high voltage architecture of the electrified power train. A data model is set up to describe the architectures in an accessible numerical structure. A combinatorial algorithm is presented, which is able to construct new architecture variants and which is based on a numerical implementation of the used developing rules. With the ability to evaluate architectures against multiple criteria, an evolutionary optimization algorithm generates the optimized architectural concept. Keywords: Automotive, Combinatorial Optimization, Data Structure, Decision Making Methodology, Electric Power Train, Electric/Electronic-Architecture, Evolutionary Optimization, Multi-Criteria Evaluation. 343 MO4D (invited) Surrogate-Based Algorithm for Computing an Upper Bound set for the 0/1 Bi-Objective Bi-Dimensional Knapsack Problem Cerqueus, Audrey (*) University of Nantes Gandibleux, Xavier University of Nantes Przybylski, Anthony University of Nantes In this work, we consider the binary knapsack problem with two objectives and two dimensions. Our purpose is to compute a tight upper bound set for the set of nondominated points of this problem. To do this we consider the surrogate relaxation. Its principle is to aggregate the constraints of the problem according to a multiplier u. Thus we obtain a single dimensional problem. In the single-objective case, the solution of this relaxed problem allows to obtain an upper bound whose quality relies on the choice of the multiplier. The problem consisting of finding the tighest possible bound is called dual surrogate problem. With two objectives, we propose to solve the convex relaxation of the surrogate relaxation. It gives a upper bound set for the 2O2DKP. The upper bound sets obtained with different multipliers are not necessarily comparable. However the intersection of the search spaces induced by those bound sets defines a tighter upper bound set. We propose a generalization of the dual surrogate problem for the bi-objective case. We call it the optimal convex surrogate (OCS) upper bound set. This upper bound set can be obtained using all the possible multipliers. Thus this is the tighest possible upper bound set based on the convex relaxation of the surrogate relaxation for the 2O2DKP. Even if the number of possible multipliers is infinite, the number of different upper bound sets is finite. Properties of the different multipliers u and dominance relation between upper bound sets are studied. An exact and an approximative method to compute the OCS upper bound set are derived. We tested numerically those two methods on a 2O2DKP benchmark and we compare them a state-of-the-art approximative method. The results obtained are encouraging for the use of this bound set computation in an exact solving method for the 2O2DKP. Keywords: Bound Sets, Multi-Dimensional Knapsack Problem, Multi-Objective Optimization, Surrogate Relaxation. 344 TU3B (invited) Sustainability Assessment and Spatial ANP: a Methodological Proposal for Studying the New Megacity Region Turin-Milan Ferretti, Valentina (*) Bottero, Marta Politecnico di Torino Mondini, Giulio Sustainability is a multidimensional concept that includes socio-economic, ecological, technical and ethical perspectives and thus leads to issues simultaneously characterized by a high degree of conflict, complexity and uncertainty. When speaking about sustainability in urban and territorial planning, decision-making thus requires the consideration of trade-offs between many, often conflicting, objectives. In this context, it is generally agreed that Multicriteria Decision Analysis (MCDA) can offer a formal methodology to deal with partial availability of information and conflicting objectives. Given the importance of the spatial dimension in territory-related decision-making problems, there is a growing interest in the integration of Geographic Information Systems and MCDA methods. The result of this integration, known as Spatial Multicriteria Evaluation, provides a tool able to support both the generation of alternatives and their comparison, taking into account the spatial distribution of each criterion, as well as the Decision Makers’ preferences. This contribution considers the problem of sustainability assessment in urban and territorial planning projects by making use of the spatial ANP approach. In particular, starting from the real challenge represented by the growing megacity region Turin-Milan in Northern Italy, the study explores the contribution of the spatial ANP for structuring the decision-making problem under analysis, paying particular attention to the elicitation of the priorities and taking into account the opinions of different experts in the evaluation process. Moreover, the analysis highlights opportunities and risks of the metropolitan region in order to support the definition of future development scenarios. This research has an innovative value, since very few experimentations of the spatial ANP exist in this decision-making context. Keywords: Analytic Network Process, Decision Support, Environmental Decision Making, Historical Industrial Buildings, Spatial Multi-Criteria Evaluation, Sustainability Assessment. 345 TH4C (contributed) Sustainable National Transport Planning – Governance with MCDA? Jensen, Anders Vestergaard (*) Leleur, Steen Technical University of Denmark Zietsman, Joe Texas A&M University System Planning for a national sustainable transport system is a complex task and with a conflict potential since it involves taking into account a wide range of criteria. In assessment of public policies, multi-criteria decision analysis (MCDA) seems adequate, since it facilitates the use of both qualitative as well as quantitative measurement scales, which makes it possible to address multidisciplinary problems. MCDA has seen a widespread decisionsupport function in public decision making in sectors such as energy and environment. Over the last decade transport planning, similar to other sector planning, has undergone a change from traditional planning to governance. This together with an increasing emphasis on sustainable development calls for new procedures, institutions and planning tools. Unlike cost-benefit analysis and environmental impact analysis, MCDA is rarely required by national laws or directives. Nonetheless, examples of public support of MCDA can be found: An EU guide points out that MCDA facilitates the participation of all actors and helps reaching a compromise or defining a coalition of views. UK has put a focus on MCDA by publishing a general MCDA guide for official use. Italian law states that MCDA is required as regards project selection for public works. Several examples of use of MCDA in the public domain exist for e.g. Sweden, Portugal and France. This paper evaluates the use of MCDA in national transport planning by its strengths and weaknesses in assessing the impacts of public policy options up for examination within sustainable national transport planning. The evaluation is based on a review of identified relevant literature. The overall key issue of this paper is to highlight the application potential of MCDA in national transport planning and its utility to policymakers engaged in such planning for a sustainable development of the transport system. Keywords: Governance, National Transport Planning, Public Policy, Sustainability, Transport Policy. 346 WE1E (invited) Sustainable Supply Chains: Contributions from MCDM Karpak, Birsen (*) Youngstown State University Globalization places demands on supply chain management beyond pure economic issues. Cost minimization, revenue or profit maximization by itself are not sufficient; for example: fair labor conditions, diversity, safety, and environmentally friendly product development (greener product design) and production (cleaner process technology) are also important and therefore sustainable supply chains are inherently multiple criteria problems. In this research we explore multiple criteria approaches to sustainable supply chains. Different multiple criteria methodologies used, time distribution of the articles, journals in which these articles published will be given. Most of the articles found in literature were case based. It is understandable since we need to understand the problem in real setting. This can be explained by the fact that the sustainability area is a relatively new research field and researchers need to do more case study work to understand the real issues and problems, something that case study methodology is well-suited for. Based on this literature search we found that, still environmental dimension dominates, social aspects are widely ignored, and social criteria need further exploration. Our overall conclusion is that sustainable supply chains are prosperous area for MCDM community. Keywords: Multi-Criteria Decision Making, Sustainable Supply Chains. 347 MO2B (contributed) Synhronous Usage of Parameterized Achievement Scalarizing Functions in Interactive Compromise Programming Nikulin, Yury (*) University of Turku Karelkina, Volha University of Turku This research addresses interactive multicriteria optimization approach based on parameterized achievement scalarizing function. We introduce a method of reflecting the decision maker’s preferences by means of changing the weights in the achievement scalarizing functions. A decision making process is simulated for the three-objective median location problem. Preliminary computational experiments illustrate the way how synchronous usage of achievement scalarizing functions may reduce the number of iterations in an interactive process and speed up its convergence to the most preferred solution. Keywords: Achievement Scalarizing Functions, Interactive Methods, Location, Parametrization, Synhronous Methods. 348 MO2G (invited) Synthetic Indicators of Mutual Funds’ Environmental Responsibility: An application of the Reference Point Method P´erez-Gladish, Blanca (*) University of Oviedo M´endez Rodr´ıguez, Paz University of Oviedo Cabello, Jose Manuel University of M´alaga Ruiz, Francisco University of M´alaga Socially Responsible Investing (SRI) is broadly defined as an investment process that integrates not only financial but also social, environmental, and ethical (SEE) considerations into investment decision making. SRI has grown rapidly around the world in the last decades. In the last years, given the causes of the 2008 financial crisis, ethical, social, environmental and governance concerns have became even more relevant investment decision criteria. However, while a diverse set of models have been developed to support investment decision-making based on financial criteria, models including also socially responsible and/or environmental responsibility criteria are rather scarce. The aim of this paper, in which we focus on the environmental dimension, is to assist individual investors in their investment decisions providing them with a synthetic indicator of mutual funds’ environmental responsibility, which is by nature a multicriteria concept and therefore multicriteria techniques are to be used to measure it. The proposed approach is based on the double (reservation-aspiration) reference point method. This scheme is applied to each of a set of US equity mutual funds’ randomly selected, in order to determine, on the basis of a given set of indicators, a pair of synthetic indicators that measure the weak and the strong environmental responsibility degree of each mutual fund relying on the particular preferences of the Decision Maker. Keywords: Environment, Equity Mutual Funds, Finance, Reference Point, Socially Responsible Investment, Synthetic indicators. 349 MO5H (poster) Systemic Decision Making: A New Holistic Approach in Ahp-Multiactor Decision Making Altuzarra, Alfredo (*) University of Zaragoza Gargallo, Pilar University of Zaragoza Moreno-Jim´enez, Jos´e Mar´ıa University of Zaragoza Salvador, Manuel University of Zaragoza Systemic decision making is a new approach to dealing with complex multiactor decision making problems in which their individual preferences over a fixed set of alternatives are incorporated in a holistic view under the “principle of tolerance”. This new approach integrates all of these preferences even if they are encapsulated in different “individual theoretical models”, the only requirement being that they must be expressed as some kind of probability distribution. In this paper, assuming that the Analytic Hierarchy Process (AHP) is the multicriteria technique employed to rank alternatives and that a Bayesian methodology is used to calculate the priorities of the alternatives, we present a new methodology to integrate the individual visions of the reality. To that aim we define a tolerance distribution which allows the incorporation into the model of the preferences of all the actors involved in the resolution process. A mathematical justification of this distribution, a study of its statistical properties and a Monte Carlo Method to draw samples are provided. Finally, the application of the methodology to the AHP-multiplicative model with lognormal errors and some illustrative examples are included. Keywords: Analytic Hierarchy Process, Bayesian Inference, Cognitive Perspective, MultiActor Multi-Criteria Analysis, Principle Of Tolerance, SIR, Systemic Decision Making, Tolerance Distribution. 350 TH2B (contributed) The Bi-Criterion Adaptive Stochastic Knapsack Problem Andersen, Kim Allan (*) Aarhus University Nielsen, Lars Relund Aarhus University Ehrgott, Matthias Lancaster University Pretolani, Daniele UNIMORE We have a known capacity of a resource, and a set of projects. Each project requires some units of the resource which is given by a discrete probability distribution. The resource requirements become known when a project has been selected. In that case two rewards are received, which only depend on the project chosen. The goal is to design a set of resource adaptive strategies for choosing the projects such that the total expected value of the two objective functions is maximized. vskip 0.3 cm We describe a two-phase method for solving the problem, based on the hypergraph concept. Preliminary computational results are presented. Keywords: Adaptive Strategies, Bicriteria Model, Stochastic Knapsack. 351 TU3C (contributed) The Decision Making Model of Service Improvements of Smart Home System based on IOA-NRM Approach Huang, Chung-An (*) National Chiao Tung University Internet applications are becoming more and more popular and the development of information and communication technologies and cloud database substantially change the lifestyle of modern family. Chip technology industry is leading smart home as a new blue ocean market of computer network applications. In recent years, SHSs (Smart home system) are more and more popular due to internet penetration and the development of private cloud system. The SHSs offer a convenient lifestyle by integrated service platforms supporting many devices and services database. This study tries to discuss the driving forces of the value of SHSs. The issue of SHS has become a trend and researchers began to put emphasis on the future application in SHSs and how the SHSs change the users’ daily life. Under the trend of smart home market prices gradually decline, multifunction computer home users gradually increase, as well as the growth number of broadband Internet home users, SHSs (Smart Home System) experience rapid development and diffusion. The main function of current service in SHSs is to share files as well as sharing broadband Internet resources between family users. Future multimedia entertainment applications are expected to be able to cause family consumers being interested in smart home system products, which is currently the main purpose of smart home industry. The traditional MCDM (Multiple criteria decision making) evaluation model does not consider the structural ration of the aspect/criteria of evaluation. However, the structural ration problem still exists in the decision system. As a result, some researchers have developed a new structural ration technique to improve upon the original MCDM model. This study develops a novel MCDM approach to solve the structural ration problem. We demonstrate the effectiveness of the model by applying empirical location-based services cases. Keywords: DEMATEL, Network Ration Map, Smart Home System. 352 TH4C (contributed) The Derivation of Weights for Sustainability Criteria: A Framework for Corporate Assessment in Sugar Manufacturing Sureeyatanapas, Panitas (*) The University of Manchester Bamford, David University of Huddersfield Yang, Jian-Bo The University of Manchester In this study, the Evidential Reasoning (ER) approach is applied in assessing corporate sustainability performance. By employing the ER theory, a robust self-assessment model can be developed, which is able to logically integrate both quantitative and qualitative criteria and to capture various forms of uncertainties in subjective evaluation. As the ER theory assumes that each criterion plays a role which is equal to its weight, a logical procedure to justify and investigate weights is important. The objective of this paper is to investigate the importance of each sustainability criterion by focusing on the case of sugar manufacturing companies. The importance of each criterion, defined as its power in discriminating different companies in terms of their ability to maintain businesses in the long run, is analysed by means of both relative weight and its real power. The latter is considered from the assumption that the degree of the real power of each criterion might be higher or lower than its relative weight, particularly when an overlap among several criteria exists. The weights are investigated by means of a questionnaire and interviews. The respondents of both methods are the managements of sugar manufacturing companies in Thailand. The weights of criteria lying on the same branch of the hierarchy are analysed relatively. The weight vectors deduced from the interviews are combined based on the Weighted Geometric Mean (WGM) method. The results are cross-compared with the weight intervals determined from the questionnaires. The possibility of the difference between the relative weight and the actual weight of each criterion is deduced based on statistical analysis. The weight information can be used to support complex decision-making processes such as corporate sustainability assessment. The findings are also meaningful in the sense that the relative weights may not always precisely explain actual contributions of the criteria toward the overall objective. Keywords: Criteria, Multi-Criteria Decision Analysis, Multi-Criteria Decision Making, Performance Measurement, Sustainability, Sustainability Assessment. 353 WE2C (contributed) The Evaluation of Competitiveness of Visegrad Four NUTS 2 Regions by the Hybrid Eigenvalue-Fuzzy Cognitive Map Approach Kiszova, Zuzana (*) Silesian University in Opava Mazurek, Jiˇr´ı Silesian University in Opava Nevima, Jan Silesian University in Opava The world of tomorrow is the world of the most competitive ones, and this concerns individuals as well as enterprises, regions or whole countries or unions of countries. The aim of this article is to evaluate competitiveness of Visegrad four NUTS 2 regions by the hybrid eigenvalue-fuzzy cognitive map approach developed by Mazurek and Kiszova (2012), which is an MCDM alternative to the standard analytic network process by T. L. Saaty. Regions were compared by the following criteria: gross domestic product (GDP), gross fixed capital formation (GFCF), net disposable income (NDI), employment rate (ER), knowledge intensive services (KIS) and patents (PAT) during programming period of years 2000 – 2006. The most competitive regions were found as well as the least competitive ones. These findings can be useful especially for the latter regions, which can adopt measures (collaboration, investment, benchmarking, etc.) necessary to catch up with the more competitive regions, and also for policy makers, as regional or national governments might stimulate the development of the least competitive regions e.g. from European structural or cohesion funds. Keywords: Competitiveness, Fuzzy Cognitive Map, Hybrid Eigenvalue, Multi-Criteria Decision Making, Regions, Visegrad Four. 354 MO3F (contributed) The Global WASF-GA: An Evolutionary Algorithm in Multiobjective Optimization to Approximate the whole Pareto Optimal Front Ruiz, Ana Bel´en (*) University of M´alaga Luque, Mariano University of Malaga Saborido-Infantes, Rub´en University of M´alaga In this paper, we propose a new evolutionary algorithm for multiobjective optimization called the Global Weighting Achievement Scalarizing Function Genetic Algorithm (Global WASF-GA), which is based on the preference-based EMO algorithm called WASF-GA. The main purpose of the new algorithm is to approximate the whole Pareto Optimal front, for what an achievement scalarizing function (ASF) based on the Tchebychev distance is used as the fitness function. This ASF depends on a reference point, which contains desirable objective function values, and on a vector of weights. In our algorithm, an utopian objective vector is used as the reference point, and we have considered a set of weight vectors which are as well-distributed as possible in the weight vector space. Basically, the main idea of the Global WASF-GA is to approximate the whole Pareto optimal front by projecting the utopian vector using this ASF and varying the vector of weights in the whole weight vector space in order to cover the whole Pareto Optimal front. For that purpose, at each generation, the individuals in the current population of parents and offspring are classified into different fronts in the following way: each front is formed by the solutions with the best values of the ASF for each one of the weight vectors in the sample. When a front is obtained, the solutions which has been included in that front are removed temporarily from the population and, subsequently, the next front is obtained in the same way. This process is repeated until every individual has been included in some front. The fact of using well-distributed weight vectors and varying them in the ASF, as well as considering an utopian vector (which dominates the ideal vector) as the reference point, enables us to obtain a final set of nondominated solutions that can approximate the whole Pareto front. Keywords: Achievement Scalarizing Functions, Evolutionary Multi-Objective Optimization, Multi-Objective Optimization, Pareto Optimal Solutions. 355 TU4B (contributed) The Optimal Resort Development Investment Projects with a Replacement Cost Perspective Chang, Chiao-Chen (*) National Dong Hwa University Lin, T. Tyrone National Dong Hwa University While the past research on the topic has generally examined the resort and theme park projects related to modeling the project decision-making and investment schedule optimization, numerous questions exist about why companies choose to invest in the resort and theme park development projects. We address the concept of comparing a complete cycle of replacement cost by providing a new model as it relates to follow a random Poisson distribution under the traditional calculation method of the net present value in different projects. The modeling results performing the model calculations show that choosing the full region type project is the optimal resort development project. However, a sensitivity analysis for several parameters such as the initial investment, the per unit time of the average customer arrival rate, the per unit time of the expected variable cost for a person, and the risk premium adjusted discounted rate is also conducted in this paper. These findings have important implications to the leisure investment companies alike as greater efforts are made to choose the optimal resort and theme park development projects. Keywords: Decision Analysis, Net Present Value, Replacement Cost, Resort Development Investment, Stochastic. 356 TU2C (contributed) The Relative Importance of the Criteria and the Ratio of the Directional Derivatives Sarkisyan, Rafael (*) Eskova, Olga Moscow University of Engineering The construction of effective interactive algorithms and procedures for solving applied multiple criteria problems of planning, projecting and management inevitably makes it necessary to take into account informal information about the relative importance of the criteria themselves. The schemes proposed by A. Geoffrion and his school back in 1970ies, based on the concept of “indifference curves” and “ideal proportions”, in principle allow to solve a multiple criteria problem with the help of convergent man-machine algorithms, provided the criteria functions are continuous, and the set of solutions, convex and compact. The mechanism of the “ideal proportions” is based on the replacement of the relative importance of criteria ωij = (∂u/fi )/(∂u/fj ) by the ratios ωij = δi /δj , i, j = 1, . . . , m, where δi and δj are “ideal proportions” along the direction of steepest ascent of the utility function u(f ) in the plane (fi , fj ), the values of the other criteria being fixed. The quantities δj can be expressed through the direction e of steepest ascent of the utility function in the set of solutions D as δj ≈ fj (xk + σe) − fj (xk ), j = 1, . . . , m, where xk is the current point, σ is the step magnitude, such that xk + σe ∈ D, σ ∈ (0, ρ), ρ 0. Decomposing the function δj (σ) into series and using the expression for the derivative of differentiable directional functions dfj (xk )/de = limσ→0+ (fj (xk + σe) − fj (xk ))/σ = S jT e, where S j is the gradient of the function fj (x), j = 1, . . . , m, for ωij one can construct the ratio ωij = S iT e/S jT e, i, j = 1, . . . , m. For the pair of target functions (fi (x), fj (x)) they connect the ratio of subjective estimates δi and δj with the ratio of directional derivatives S iT e/S jT e, i, j = 1, . . . , m. These ratios have a number of effective analytical properties; they can be used as a criterion of coordination and correction of quantities δi and δj , as well as for constructing forecasts regarding the preferred directions of motion in the set D and, finally, for expanding the conception of methods of possible directions of Zoutendijk as applied to the multiple criteria problems. Keywords: Ideal Proportions, Indifference Curve, Man-Machine Algorithm, Multi-Objective Optimization, Ratio Of Directional Derivatives, Relative Importance. 357 WE2B (contributed) The SMAA-PROMETHEE Methods Corrente, Salvatore (*) University of Catania Figueira, Jose Technical University of Lisbon Greco, Salvatore University of Catania We apply the Stochastic Multiobjective Acceptability Analysis (SMAA) to the classical PROMETHEE methods and to the Bipolar PROMETHEE methods. Bipolar PROMETHEE methods extend the classical PROMETHEE methods to the case of interacting criteria considering the bipolar approach to Multiple Criteria Decision Aiding (MCDA) and aggregating positive and negative preferences by means of the bipolar Choquet integral. SMAA explores the whole set of parameters compatible with some preference information provided by the DM taking also into account imprecision on the evaluation of alternatives on considered criteria. We compare the results given by SMAA approach with the results supplied by Robust Ordinal Regression (ROR) that has been recently proposed to derive robust conclusions through the use of the concepts of possible and necessary preferences. Keywords: Bipolar Method, Interaction Among Criteria, PROMETHEE, Robust Ordinal Regression, SMAA. 358 TH3D (contributed) The Support of Decision Criteria Weights Elicitation based on their Probabilistic Characteristics Brzostowski, Jakub (*) Silesian University of Technology Roszkowska, Ewa University of Bialystok Wachowicz, Tomasz University of Economics in Katowice Based on the data gathered by the Negotiation Support System - Inspire we analyze the probabilistic patterns occurring in the collective preferences of large group of decisionmakers preparing to conduct negotiations in the stage of their pre-negotiation phase. The behavior of negotiators in the preparation phase indicates on some regularities in the distributions of vectors of criteria weights that allows to form a probabilistic characteristic of these crucial preferences’ elements. After performing the normality tests we discovered that if the issues (criteria) are ordered from the most to the least important than the pairs of consecutive issue weights are normally distributed. Such an observation allows us to use the mean values, standard deviations, and correlation levels to describe the preferences of a population of decision-makers. The final result of our research is a decision supporting tool facilitating the process of decision criteria weights elicitation by allowing the decision-maker to have insight into collective preferences. Such a supportive tool may be used in variety of decision problems in the preliminary step of analysis followed by the use of some of the MCDM methods to conduct the next steps of preferences’ analysis. The illustration of marginal and conditional probability distributions over the issue weights allows to gain an intuition about global preferences and to manipulate the potential issue weights resulting with an indication of the part of population having similar preferences. Any manipulations with particular issue weight result in the modification of conditional distributions corresponding to the remaining issues since there is significant correlation between different issues related to the decision problem. The aim of the proposed system is the indication how the adjusted preferences of the decision-maker relate to the global, synthetic preferences of a population of decision-makers solving similar problems in the past. Keywords: Applied Probability, Decision Maker’s Preferences, Decision Support Systems. 359 TU2G (invited) The Treatment of Actuator Failures in Mechatronic Systems by Multi-objective Optimization Methods Horenkamp, Christian (*) University of Paderborn Dellnitz, Michael University of Paderborn Ober-Bl¨obaum, Sina University of Paderborn Timmermann, Robert University of Paderborn Many modern mechatronic systems are desired to be optimal with respect to several concurrent objectives such as energy and comfort. There exist many well-known multiobjective optimization methods for the numerical computation of the so-called Pareto optimal configurations. In this talk we consider actuator failures in mechatronic systems. Generally, in case of an actuator failure, the once chosen working point will not be Pareto optimal anymore. Therefore, a new optimal configuration has to be calculated. Additionally, the new working point should be as close as possible to the initial point in objective space. Classical optimization approaches would calculate the whole set of Pareto optimal configurations for every possible failure case already during the development of the system. This is numerically very costly in systems with a high number of actuators and therefore often not practicable. We present a method which is invoked at the occurrence of an actuator failure and cal-culates a new Pareto optimal configuration during runtime of the system without explicitly calculating a whole set of Pareto optimal configurations. This method is based on path-following techniques for parameter-dependent multi-objective optimization problems and will be illustrated by the active suspension module of a railbounded vehicle. Keywords: Actuator Failure, Parameter-Dependent Problem, Path-Following Methods. 360 MO3F (contributed) Theoretically Analyzing Simple Interactive Evolutionary Multiobjective Optimizers Brockhoff, Dimo (*) INRIA Lille - Nord Europe Evolutionary Multiobjective Optimization algorithms (EMOAs) typically aim at finding an approximation of the entire Pareto front. Interactive EMOAs, however, have been proposed in recent years to focus the search only on regions of the front in which a decision maker (DM) is interested and which she/he is defining by revealing information about preferred solutions successively during the search. While several interactive EMOAs are available, it is difficult to compare the approaches and to make some recommendations about which method to use in practice. Moreover, a thorough theoretical understanding of their runtime behavior is lacking which, if available, would assist in choosing an algorithm over a clearly outperformed one. In this talk, I will briefly present the first results on theoretically proving bounds on the expected number of function evaluations and number of DM interactions until the most preferred solution of the DM is found. The results hold under the assumption that the DM can choose the preferred solution in a pair of incomparable points based on her/his internal utility function. While the first results for two simple interactive EMOAs on two pseudo-boolean test functions have been published at PPSN’2012, I will discuss in this talk new results on how the theoretical bounds change when the interactive EMOAs are allowed to learn the utility function during the search. Keywords: Evolutionary Multi-Objective Optimization, Interactive EMO, Learning Utility Functions, Pseudo-Boolean Search Space, Runtime Analysis. 361 TH2G (invited) Transforming Constraints into Objectives: Experiments with Bidimensional Knapsack Problems Schulze, Britta (*) University of Wuppertal Klamroth, Kathrin University of Wuppertal Paquete, Luis University of Coimbra We consider constrained combinatorial optimization problems and formulate associated multiple objective optimization problems, where one or several of the constraints are relaxed and interpreted as additional objective functions. In this way, the trade-off between constraint satisfaction on one hand and original objective value on the other hand can be analyzed. As a concrete example problem, we consider bidimensional knapsack problems (i.e., one objective and two knapsack constraints) and their associated biobjective, singleconstraint knapsack problems. It is shown that a dynamic programming based solution approach can be adapted in such a way that a representation of the nondominated set is obtained at little extra cost as compared to the solution of the original problem. In this context we discuss strategies for bound computation and for handling negative cost coefficients, which occur through the transformation. Numerical results comparing the single- and multiple objective approaches are presented. Keywords: Bidimensional Knapsack Problem, Constraint Handling, Dynamic Programming, Multi-Criteria Combinatorial Optimization. 362 MO2G (invited) Tri-criterion Inverse Portfolio Optimization with Application to Socially Responsible Mutual Funds Wimmer, Maximilian (*) University of Regensburg Steuer, Ralph E. University of Georgia Hirschberger, Markus Munich Re Utz, Sebastian University of Regensburg We present a framework for inverse optimization in a Markowitz portfolio model that is extended to include a third criterion. This causes the traditional nondominated frontier to become a surface. Until recently, it had not been possible to compute such surfaces. But by using a new method that is able to generate the nondominated surfaces of tricriterion portfolio selection problems, we are able to compute via inverse optimization the implied risk tolerances of given funds that pursue an additional objective besides risk and return. In applying this capability to a broad sample of conventional and socially responsible mutual funds, we find that after the screening process there appears to be no significant difference in how assets are allocated in socially responsible and conventional mutual funds. Keywords: Inverse Optimization, Multi-Criteria Decision Making, Multi-Objective Optimization, Non-Dominated Surfaces, Portfolio Selection, Socially Responsible Investment. 363 MO4E (contributed) Triple Bottomline, Hyper-radial-Visualisation-Based ’Decision-Making by Shopping’ for a Land Use Management Problem using Evolutionary Multi-objective Optimisation Chikumbo, Oliver (*) Scion Research Deb, Kalyanmoy Michigan State University Goodman, Erik BEACON – NSF Center for the Study of Evolution in Action A land use multi-objective optimisation problem for a 1500ha farm with 315 paddocks was formulated with 14 objectives (maximising sawlog production, pulpwood production, milksolids, beef, sheep meat, wool, carbon sequestration, water production, income, and Earnings Before Interest and Tax (EBIT), and minimising costs, nitrate leaching, phosphorus loss, and sedimentation). It involved 111 land use options for most of 315 paddocks and spatial constraints that made the optimization problem combinatorial. Each point in the search space represented a set of land-use management actions taken within a 10-year period and their related management options spanning a planning period of 50 years. Our aim was to develop a procedure that would reduce the cardinality of considered solutions through a transparent decision making process involving many-objective optimization and multiple heterogeneous stakeholders in order to address the Triple Bottomline of balancing environmental, economic and social aspects. The optimization was achieved by using a modified Reference-point-based Non-dominated Sorting Genetic Algorithm (R-NSGA-II) augmented by simulated epigenetic operations. The decision making process was iterative and multi-staged. The first stage involved preference construction using virtual reality (VR)-based visual steering achieved through a Hyper Radial Visualisation (HRV) process that took uncertainties into account. After narrowing down to a few trade-off solutions from the first stage, decision makers in the second stage used AHP to arrive at a single compromise solution. Highlights of the overall approach are the development of an innovative epigenetics-based multi-objective optimiser, uncertainty incorporation in the search space data, and a multi-stage decision-making on a multi-dimensional space through a VR-simulation-based visual steering process controlled at its core by MCDM-based processes. The proposed approach has widespread applicability to many other ”wicked” societal problems of interest. Keywords: Evolutionary Algorithm, Hyper Radial Visualisation, Hyperspace Pareto Frontier, Multiplicative Analytic Hierarchy Process, Reference Point based NSGA-II, Triple Bottomline, Visual Steering. 364 FR2E (invited) Two-Phase Method for Multiobjective Mixed 0-1 Linear Programming Vincent, Thomas (*) University of Nantes Gandibleux, Xavier University of Nantes Przybylski, Anthony University of Nantes Many concrete and important problems can be formulated by a mixed-integer linear programme. For these problems, several conflicting objectives can have to be taken into account, leading to a multiobjective programme. Mavrotas and Diakoulaki (1998,2005) have proposed a branch and bound method aimed at solving multiple objective binary mixed integer programmes. In the original paper, some dominated points may remain in the output of the algorithm and this issue was only partly addressed in the latter. Vincent et al (2013) proposed a corrected method based on an appropriate representation of the nondominated set. In fact, continuous sets of nondominated points must be taken into account in addition to discrete points. Those corrections were only intended for biobjective problems, thus it remains to handle three and more objectives. In this talk, we will first discuss the difficulties when three objective functions are considered. Besides the expected loss of the natural order for nondominated points, some others are typical of mixed integer programmes. For instance, a face filtered by dominance may be not convex. We propose an adapted representation of the nondominated set for the three-objective case with an update procedure. A branch and bound algorithm can next be applied. Finally, we consider an adaptation of the two-phase method that was already applied for the biobjective case. The difficulty lies here in the determination and the efficient partitioning of the search area. Indeed, if in the combinatorial case, the search area can be defined only using corner points, here corner edges are also necessary. We propose a projective algorithm to update the description of the search area for the mixed-integer case using all known feasible points (and faces). This algorithm is a generalization of (Przybylski et al. 2010). The proposed algorithms are tested on various instances and their results are discussed and compared. Keywords: Multi-Objective Mixed Integer Programming, Two-Phase Method. 365 TH4E (contributed) Unsupervised Update of the Preferences on Numerical and Multi-Valued Categorical Criteria from the Analysis of the User Behaviour Valls, Aida (*) Isern, David Marin, Lucas Moreno, Antonio Universitat Rovira i Virgili The quality of multi-criteria decision support systems (MCDSS) depends heavily on their knowledge about the user preferences on the values of numerical and categorical attributes. There are many scenarios in which MCDSS are continuously interacting with the user (e.g. filtering news from media or from social networks), and the feedback obtained from the users (e.g. the news they actually read) can be used to learn their preferences implicitly. In our previous works it was shown how the selections of the user could be leveraged to effectively learn a symmetric triangular utility function, whose maximum was the most preferred value. For categorical attributes, each possible value was given a linguistic preference score. In this work we present two improvements to the learning method. For numerical attributes, the utility function has been generalised, including five parameters: preferred value, right/left intervals and right/left slopes. In this way the utility function is much more expressive and it is possible to deal with a wider set of scenarios (e.g. taking 3 as the preferred value, the preference may decrease quickly below 3, but reduce more slowly over 3). The MCDSS analyses carefully the user continuous selections to learn these parameters for each numerical attribute. The second improvement allows alternatives to take multiple values in the categorical attributes. The presence of multiple values makes it more difficult to analyse the user choices and to infer why the user has chosen (or discarded) a certain alternative. The profile learning method can now deal with this kind of scenarios, being able to assign efficiently a qualitative preference score to each possible value. This preference learning method has been tested for recommending restaurants in Barcelona, obtaining encouraging results. This work has been supported by a pre-doctoral grant from University Rovira i Virgili and the Spanish projects DAMASK (TIN2009-11005) and SHADE (TIN2012-34369). Keywords: Decision Support Systems, Multi-Criteria Decision Making, Preference Learning. 366 MO5H (poster) Use of Fuzzy ANP-VIKOR Technique in Multi-Criteria Decision-Making of Satisfaction Evaluation for Business-to-Customer (B2C) Services Improvement Chern , Yuching (*) National Chiao Tung University Tzeng, Gwo-Hshiung Kainan University Customer satisfaction in e-commerce (e-satisfaction) is a key determinant that promotes the repurchase actions at Website stores and creates marketing values for business-toconsumer (B2C) service firms. However, while the measured values of human perceptions towards various criteria are often vague data, most conventional measurement methods can not precisely handle imprecise numerical quantities. Therefore, this paper aims to present a fuzzy multiple-criteria decision-making model with fuzzy ANP-VIKOR approach to manage the assessment problem of B2C service e-satisfaction measures. First, the analytic network process (ANP) is applied for determing weights of essential factors that enhance B2C service e-satisfaction. Secondly, the logic of triangular fuzzy numbers are used in developing the fuzzy VIKOR algorithm for prioritization among the alternative B2C services providers and for identifying the satisfaction gaps between alternatives and the aspired-level, specifically intended to reduce gaps for B2C service improvement. An empirical case illustrates the applicability and effectiveness of the proposed fuzzy ANPVIKOR method, and the result by the TOPSIS method is compared. The multi-criteria consumers’ e-satisfaction measures derived from this study can further support decisionmakers in service contents selection, and indicate solutions to enable customer satisfaction with B2C services improvement. Keywords: Analytic Network Process, E-Satisfaction, Fuzzy Logic, Multi-Criteria Decision Making, VIKOR. 367 FR2D (contributed) Use of Multiattribute Utility Function for a Cost-Utility Analysis Kangaroo Mother versus “Traditional” Care for Premature Infants in Bogota, Colombia Castillo, Mario (*) Bernal, Astrid R´ıos, John-Jairo Universidad de los Andes Charpak, Nathalie Kangaroo Foundation C´ordoba, Mar´ıa-Adelaida Ruiz, Juan-Gabriel Universidad Javeriana Trujillo, Tammy During last years, the high impact of prematurity on mortality and morbidity of newborns has been identified. Currently, the standard care given to premature newborns is to keep the newborn hospitalized in a neonatal unit. Another alternative is the Kangaroo Mother Program (KMP) which aims to obtain a significant reduction in hospitalization costs of newborns and nosocomial infections, among others. A cost-utility analysis (CUA) of alternatives was performed, in which utility of the treatments was estimated with a multicriteria model. As a part of CUA, a methodology was designed to evaluate utility of medical treatments using MAUT approach, which was applied specifically to treatment of premature newborns in Bogota, Colombia. The methodology starts constructing a multiattribute utility function. Next, it was defined the weight of each variable, and utility of each states. Finally, the average utility of newborns in each treatment was estimated. Currently, cost estimation is in process and is not included in this presentation. Measuring utility was established with the participation of a group of medical experts in pediatrics. The models and the analysis were constructed based on the study results Charpack, Ruiz, Calume, & Charpack, 1997, that built a database of 593 premature newborns. A second model was built, in which mutually exclusive health states were defined and scored, and it was used to validate results. The main result of the implementation of the methodology was the construction of a model for evaluating the utility of the newborns. The multicriteria model results showed statistically significant difference between treatments in favor of KMP. Similar results were obtained by the second utility model. The multiattribute utility function constructed enabled to evaluate utility of newborns health state, and compare both treatments analyzed. It allowed capturing judgments of experts in a rigorous way. Keywords: Cost Utility Analysis, Cost-Effectiveness Analysis, Multi-Attribute Utility Theory, Premature Newborns Care. 368 TU3F (invited) Using Choquet Integral as Preference Model in Interactive Evolutionary Multiobjective Optimization Branke, Juergen (*) University of Warwick Corrente, Salvatore Greco, Salvatore University of Catania Slowinski, Roman Poznan University of Technology We propose an interactive multi-objective evolutionary algorithm (MOEA) that attempts to discover the most preferred part of the Pareto-optimal set. Our approach is inspired by the approach of Greenwood, Hu and D’Ambrosio (1996) which ranks a solution better than another if it is preferred by all the utility functions compatible with pair-wise preference information supplied by the user. Our approach differs from the approach by Greenwod et al. because instead of modeling user’s preferences by a linear utility function we consider a more flexible utility function - a Choquet integral. We show that our approach works well even in situations where preferences can not be expressed by a linear utility function (and the approach by Greenwood et al. fails to find the most preferred solution). In case the linear utility function can express the user’s preferences, our approach is able to get the same result as the Greenwood et al. approach, although it converges a bit slower. Keywords: Choquet Integral, Evolutionary Multi-Objective Optimization, Preference Learning. 369 TU4E (contributed) Using MACBETH with the Choquet Integral to Model Interdependencies between Indicators in the Context of Risk Management Lopes, Diana F. (*) Technical University of Lisbon Bana e Costa, Carlos Technical University of Lisbon Oliveira, M´onica Duarte University of Lisbon Risk management typically demands for comparing the consequences of different sources of risk on multiple dimensions. Multicriteria value models can assist in evaluating those consequences. Often several impact indicators should be considered in some dimensions and, therefore, modelling should identify and account for possible judgemental interactions between indicators. The Choquet Integral (CI) has been used for this purpose, namely for industrial performance aggregation, and many applications of the 2-additive CI operator have been reported in the literature, for it is argued that 2-additivity is probably the best compromise between complexity and richness of the model. In this framework, the MACBETH technique is used to determine the parameters of the Choquet model, from n+1 sets of pairwise qualitative judgements of difference in attractiveness, one for each one of the n indicators separately and another one for pairs of interdependent indicators. However, all the MACBETH judgements assessed are judgements of global attractiveness, and therefore we propose that it is more adequate to build one single MACBETH global matrix and determine all of the Choquet parameters directly from it. The proposed procedure is applied to the evaluation of the consequences of sources of risk in the risk management context at the ALSTOM Company. Keywords: Choquet Integral, Interactions, MACBETH, Risk Management. 370 TH4E (contributed) Using Methods of Multi-Criteria Decision Making to Evaluate the Sustainability of Bioenergy Concepts by Integrating Adequate Reference Points Lerche, Nils (*) University of Goettingen Geldermann, Jutta University of G¨ottingen Biomass usage for energy conversion is often associated with sustainability, as it is considered more favorable with regard to climate change than fossil fuels. Nevertheless, there are also significant concerns, particularly about the need for subsidies, the increase of monocultures or the competition for soil with food crops. Thus, it becomes apparent that the sustainability evaluation of biomass usage concepts is of multi-criteria nature and should not only imply economic, but also environmental, social and technical aspects. Hence, methods of Multi-Criteria Decision Analysis (MCDA) appear to be suitable. However, in order to evaluate different concepts with respect to sustainability, it is not sufficient just to compare several alternatives on the basis of their performance concerning different criteria. To identify suitable sustainable concepts, the integration of an adequate reference point is needed. Through this reference point it should be facilitated that a concept actually fulfills the requirements for sustainability instead of merely being better than other alternatives. Hence, the identification of a suitable reference point, which could be applied as indicator for sustainability within methods of MCDA, is crucial. Another aspect of increasing interest is the actual decision behavior of decision makers and the consideration of behavioral effects in MCDA-methods. To combine the need for a suitable reference points in combination with behavioral effects, the integration of elements of prospect theory appears to be a meaningful enhancement. Therefore, potential extensions and adjustments of existing concepts for combining the multi-criteria method PROMETHEE with prospect theory will be discussed to address the need for an adequate reference point and the integration of behavioral effects, enabling an improved assessment with respect to sustainability. In addition, a bioenergy-related example is presented to illustrate the enhancements. Keywords: Bioenergy, PROMETHEE, Prospect Theory, Reference Point, Sustainability. 371 FR3E (contributed) Using Structured Decision Making to Facilitate Environmental Risk Management Liu, Shuang (*) CSIRO Ecosystem Sciences Environmental managers and policy-makers face two major challenges when making decisions involving environmental risks. First, management decisions frequently involve trade-offs between complex and often competing environmental, social and economic consequences with potential positive or negative outcomes for different social groups. Second, understanding of these consequences is often marked by profound uncertainties. When combined, these challenges too often become excuses for maintaining the status quo instead of considering alternatives that might result in net social welfare gains. One emerging decision-aid tool that combats these two challenges by taking into account social values and uncertainty is Structured Decision Making (SDM). Compared to MCA without a public involvement component, SDM provides an opportunity for diverse stakeholder views to be explicitly incorporated within the decision-making process. In addition, the SDM can also function as a communication platform, whereby scientists, stakeholders and decision-makers can interact and discuss the uncertainties associated with environmental decisions. Thus, SDM injects scientific rigor and transparency into the decision-making process by providing an analytical structure for social complexity and an interacting forum for risk communication. In this presentation, we situate our methodology within the environmental management and MCA literature, discussing the advantages and limitations of SDM as a framework to support environmental risk management. We also discuss the potential for integrating SDM with other decision-aid tools such as decision trees, cost-benefit-analysis, and interactive gaming models. A decision support system incorporating SDM and these other tools greatly enhances the capacity of SDM in evaluating creative alternatives and addressing explicit tradeoffs associated with environmental risk management. Keywords: Biosecurity Management, Decision Support Systems, Environmental Decision Making, Group Decisions Making, Risk Communication, Risk Management, Scientific Uncertainty, Water Resources Management. 372 TH2C (contributed) Value Efficiency: A General Case Soleimani-Damaneh, Majid (*) University of Tehran & IPM Value Efficiency Analysis (VEA) is an approach for measuring the relative efficiency in DEA, Multiobjective Optimization and MCDM, incorporating the decision maker’s preferences explicitly. This technique allows us to use the value judgments in performance analysis and decision making. This paper provides new connections between different definitions of value efficiency. The importance of the (Lower semi/Upper semi) continuity of the value function is highlighted. Some problematic situations are illustrated by discontinuous examples. Also, a connection between cone-preferences and value functions is addressed. A main theorem is established providing a sensitivity-analysis-based result. It is shown that the value inefficiency with respect to a wide class of value functions can be recognized by obtaining the optimal value of an optimization problem. This class of functions contains some usual forms, including Locally Lipschitz concave functions and pseudo-concave ones. Also, we incorporate more general value functions in VEA, using the global optimality notion. Another part of this talk is devoted to nonsmooth value functions with respect to Clarke generalized gradients. Some properties of these functions are investigated. Mean-value theorem, closedness and boundedness properties of the Clarke generalized gradients set play a vital role in this investigation. Also, weightrestricted models are studied and some of their properties are addressed. Keywords: Data Envelopment Analysis, Multi-Objective Optimization, Value Function. 373 MO4C (invited) Variational-Like Inequality Problems as an Important Tool to Solve Vectorial Optimization Problems Ruiz-Garz´on, Gabriel (*) Universidad de C´adiz Batista dos Santos, Lucelina Universidade Federal do Parana Osuna-G´omez, Rafaela Universidad de Sevilla Rufi´an-Lizana, Antonio Universidad de Sevilla The purpose of this work is to investigate the relationships between solutions of Stampacchia and Minty vector variational-like inequalities and Pareto and weak Pareto solutions of vector optimization problems in Banach spaces under pseudo invexity and pseudo monotonicity hypotheses. These results enable us to relate Pareto and weak Pareto points and solutions of the weak and strong Minty, Stampacchia and perturbed vector variationallike inequalities problems. Our results extend and improve the corresponding results of others in finite dimensional spaces. Keywords: Convex Programming, Multi-Objective Optimization. 374 TH4E (contributed) Web Multicriteria Spatial Decision Support System: Integration of a Web-based ELECTRE TRI in ArcGIS Silva, Sandra (*) Instituto Polit´ecnico de Viana do Castelo Al¸cada-Almeida, Lu´ıs Dias, Luis University Coimbra Solving spatial multicriteria decision problems can benefit from both Geographic Information Systems (GIS) and Multicriteria Decision Aid (MCDA). A useful support is provided by a specific family of Spatial Decision Support Systems (SDSS) named Multicriteria Spatial Decision Support Systems (MC-SDSS), which are based on GIS and MCDA integration. MC-SDSS tools offer unique capabilities for automating, managing and analyzing spatial decision problems with large sets of feasible alternatives and multiple conflicting and incommensurate evaluation criteria. The conceptual idea in this study is based on the use of GIS capabilities to develop an adequate platform in order to use multicriteria methods available in a Web-based MCDA Server. A Web MC-SDSS methodological framework is proposed for a fully integrated system of GIS and ELECTRE TRI in ArcGIS software of ESRI that interacts with a Web Algorithms Server for MCDA techniques application. The developed Web MC-SDSS is tested on a case study analyzing environmental and socioeconomic sustainability of dairy farms in Entre-Douro-e-Minho Region. Keywords: Decision Support Systems, Multi-Criteria Decision Analysis, Outranking Methods, Spatial Multi-Criteria Evaluation, Sustainability Assessment. 375 TH4D (contributed) Web-Based Decision Support System for Solving Multiple Objective Linear Integer Problems (MOLIPs) Kirilov, Leoneed (*) Genova, Krasimira Gouliashki, Vassil Staykov, Boris Bulgarian Academy of Sciences The DSS WebOptim has the following main features: Web accessible, User-friendly, Solver-universality, Heterogeneity. The main optimization solver is the generalized scalarizing method GENS-IM. It incorporates twelve multiple objective methods. Also an evolutionary multiple objective method for generating approximate Pareto optimal solutions is available. Solving the MOLIPs by means of DSS WebOptim is very easy: 1) Each user creates personal account in the system by using simple method for web registration with username and password; 2) When the user is logged in, he/she is able to define or edit single or multi-objective optimization problems which automatically are associated with his/her personal profile; 3) To initialize solving process, an appropriate solver must be chosen. Then the problem is parsed by a specific syntax parser. After that all data is send to the corresponding solver by using inter-modular communication protocol using web service standards; 4) When the solver finds a solution or error event occurs, all data is send back to the main system, again using web service based inter-modular communication protocol. The user can preview the solution, enter new preferences or edit the problem and send it to the same or another solver for new solution. This process is continuing until the decision maker is satisfied with the current computed solution. Each solution step is kept in the system so that there is a full solution history for the problem, allowing user to go back or forth through the steps using them for start point for new solution. In conclusion, the DSS WebOptim is a very convenient tool for solving MOLIPs. The DSS WebOptim realizes the contemporary tendencies and conceptions for solving decision making problems. Acknowledgment: This research is supported in part by Bulgarian NSF, Grant No DTK02/71 and IICT-BAS research project “Applications of Operations Research for Decision Making in Engineering Systems”. Keywords: Decision Support Systems, Integer Programming, Multi-Objective Optimization, Web Architecture. 376 TU2B (contributed) What Matters to Stakeholders and Citizens in the Evaluation of Local Policies? Using MCDA to Understand Different Perspectives in Town Renewal da Silva Vieira, Ricardo (*) Antunes, Paula Universidade Nova de Lisboa Small towns provide many services to their citizens. Yet, during the last decade, there seems to be a decline on the services these towns provide. These services include hospitals, schools, transport and retail. Many policy-decisions on town renewal target behavioural changes of their citizens. In society, there are multiple legitimate values. Decisions have always a value frame behind. So, will the values used in decision-making be effective in changing people’s behaviour? If a different value set was to be considered, what alternative responses might be developed? The aim of this paper was to understand how well do the objectives/values used for the appraisal of town renewal options by policy makers represent citizens’ concerns, using Multicriteria Mapping (MCM) as a tool to elicit stakeholders values. A case study is used: Thornbury, a small town in the Southwest of England. Thornbury was experiencing a decline in retail in the town centre on the public services. Two studies were conducted: one looking into retail decline and the other to public transport decline. Individual interviews were conducted with Thornbury residents using MCM, where participants defined their own criteria (at least three) and weights. 19 participants were used for the first study and 12 participants for latter study. Some participants could only think of one or two criteria for the appraisal. Even though, results showed wide range of criteria, including qualitative criteria (e.g, being a novelty in the town, creating a purpose to come into town and Community cohesion). The range of criteria generated shows that the technical side of the problems (e.g, infrastructure or management related issues) are only a part of the issue. New criteria (e.g, Community cohesion and identity and Equity for consumers), different from criteria used by local authorities (e.g, costs of implementation, effectiveness in dealing with the issue and accessibility for retail regeneration) were deemed relevant. Keywords: Criteria, Multi-Criteria Mapping, Participatory Multi-Criteria Evaluation, Public Transport, Retailing, Town Regeneration. 377 List of Keywords 3-Objective Linear Programming Solver, 98 3PL, 59 Automated Mechanism Design, 130 Automotive, 343 Auxiliary Services Of Power Plants, 291 A* Search, 174, 274 Academic Performance Assessment, 121 Achievement Scalarizing Functions, 56, 166, 209, 316, 348, 355 Actuator Failure, 360 Adaptive Relaxation Of Constraints, 213 Adaptive Strategies, 99, 351 Additive Value Function, 278 Adjacency, 115 Adversarial Risk Analysis, 101 Aggregation Of Individual Priorities, 46 Aggregation Schemes, 108, 149, 245 Agricultural Public Goods, 120, 311 Agriculture, 311 Airport Security, 101 Allocation, 66, 306 Alternatives, 198 Amenities, 71 Analytic Hierarchy Process, 46, 57, 86, 96, 112, 127, 128, 141, 156, 172, 230, 238, 263, 266, 300, 306, 337, 338, 350 Analytic Network Process, 59, 65, 93, 94, 120, 140, 238, 303, 342, 345, 367 Animal Welfare, 146 Anytime Computing, 308 Applications, 100 Applications of MCDM, 52, 67, 70, 78, 90, 122, 145, 151, 160, 163, 194, 201, 242, 243, 251, 253, 254 Applied Probability, 359 Approximation, 124, 286 Approximation Algorithm, 285 Approximation of The Ideal Point, 85 Artificial Intelligence, 108, 170, 254, 274, 328 Aspiration and Reservation Values, 276 Assembly Line, 144 Assignment, 226, 238, 264 Auctions/Bidding, 155, 236 Augmented Goal Programming, 129 Bargaining, 136 Bayesian Analysis, 47, 208 Bayesian Inference, 182, 350 Behavioral, 51 Belief Decision Trees, 218 Belief Distribution, 182 Belief Rule-Based methodology, 108, 250 Bi-Objective Location Problem, 281 Bi-Objective Multi-Commodity Flow Problem, 48 Bi-Objective Ring Star Problem, 83 Bi-Objective Routing, 113 Bi-Objective Shortest Path Problems, 274 Bi-Objective Simplex Method, 48 Bi-Objective Traffic Assignment, 133 Bias, 134 Bicriteria Model, 52, 351 Bidimensional Knapsack Problem, 362 Bidirectional Heuristic Search, 135 Bioenergy, 119, 371 Biosecurity Management, 372 Bipolar Method, 358 Block Coordinate Descent, 164 Bound Sets, 344 Branch and Bound, 49, 286 378 Capability Based Planning, 57 Capital Budgeting, 156 Case Example, 185, 293 Case Study, 78, 94, 145, 150, 304 Category Size, 246 CD-MAMCA, 91 Characterization Of Statutory Zones, 71 Chemmical Engineering, 251 Choquet Integral, 92, 146, 199, 369, 370 City Distribution, 91 Classification, 79, 97, 138, 187, 218, 245, 279 Clinical Performance, 181 Coalition Formation, 257 Cognitive Map, 260 Cognitive Perspective, 350 Collaborative Planning, 143, 193 Combinatorial Optimization, 95, 113, 124, 190, 199, 206, 286, 316, 343 Comparative Analysis, 53, 142 Competitiveness, 61, 65, 96, 354 Composite Importance, 342 Compromise Programming, 55, 316, 337 Computational Cost, 81, 102, 253 Computationally Expensive Problems, 269 Computing Science, 172 Computing With Words, 194 Conflicting Criteria, 142 Conjoint Analysis, 186 Consistency, 300, 311 Consistency Consensus Matrix, 300 Constraint Handling, 213, 362 Constraint Programming, 287 Constructability, 71 Construction And Demolition Waste, 53 Control, 122, 170, 251, 308 Convex Programming, 374 Convex Sequences Of Weights, 313 Copula-Based Model, 183 Corporate Social Performance, 73 Correlation, 249 Corridor Location, 214 Cost Avoidance, 232 Cost Benefit Analysis, 306 Cost Utility Analysis, 101, 368 Cost-Effectiveness Analysis, 368 Costs, 325 Cournot Competition, 296 Covering Location Problem, 320 Creating Alternatives, 151 Criteria, 172, 243, 353, 377 Criterion Importance, 51, 146 Critical Paths, 79 Cross Entropy, 152 Culture, 229 Decision Analysis, 131, 134, 141, 148, 150, 159, 223, 226, 237, 302, 341, 356 Decision Maker’s Preferences, 64, 129, 149, 212, 245, 288, 304, 311, 330, 359 Decision Making, 92, 107, 122, 128, 129, 186, 240, 243, 253, 272, 279, 282 Decision Making Methodology, 343 Decision Space Diversity, 326 Decision Support, 71, 134, 155, 161, 176, 208, 302, 345 Decision Support Systems, 66, 117, 141, 172, 225, 226, 242, 273, 275, 359, 366, 372, 375, 376 Decomposition, 164 Decomposition Method, 48 DEMATEL, 65, 140, 342, 352 DEMATEL-Based ANP, 177 Dempster-Shafer Theory, 182 Differential Evolution, 203, 209 Distance-Based Methods, 157 Distributed Decision Making, 193 Distribution Centers Location Problem, 262 Dominance Measuring Methods, 82 Dominance-Based Rough Set Approach , 225, 277 Duality, 270 Durability, 268 Dynamic Approach, 167 Dynamic Network, 261 Dynamic Optimization, 251 Dynamic Programming, 331, 362 Dynamic Systems, 122, 251 Dynamic VRPTW, 256 E-Commerce, 130, 140 E-Satisfaction, 367 Earliest Arrival Flow, 261 Early Admission, 121 Econometrics, 289 Economics, 130 Ecosystems, 304 Education, 117, 289 Effectiveness, 281 Efficiency, 84, 125, 126, 281 Efficient Frontier, 307 ELECTRE Methods, 77, 222, 228 ELECTRE TRI, 71 Electric Power Train, 343 Daily Used Materials, 270 Data Envelopment Analysis, 103, 181, 283, 305, 310, 313, 314, 373 Data Mining, 66, 97, 323, 330 Data Routing, 295 Data Structure, 343 De Novo Programming, 167 Decision Aiding, 278 379 Electric/Electronic-Architecture, 343 Electricity Production Technologies, 263, 265 Elementry Education, 284 Elicitation, 149, 173, 219 Embedded Systems, 271 Empirical Study, 198 Energy Efficiency, 221, 330 Energy System, 127, 185, 240, 241 Entropy, 178 Environment, 75, 145, 162, 349 Environmental Decision Making, 70, 94, 131, 149, 208, 215, 216, 299, 323, 345, 372 Epistemology, 336 Epsilon-Constraint Method, 119 Equity Mutual Funds, 55, 126, 299, 349 Erosion Control, 210 EU, 229 Eucalyptus Plantations, 210 Evacuation, 261 Even Swaps, 134 Evidential Reasoning, 121, 182 Evolutionary Algorithm, 83, 203, 364 Evolutionary Computations, 66, 109, 294, 315, 326 Evolutionary Multi-Objective Optimization, 81, 99, 152, 168, 169, 180, 192, 196, 207, 266, 294, 315, 323, 330, 355, 361, 369 Evolutionary Optimization, 166, 343 Exergy, 239, 240 Expensive Evaluations, 130 Expert systems, 217 Exponential Family, 47 Fractional Programming, 69, 242, 333 Furniture, 338 Fuzzy Cognitive Map, 354 Fuzzy DEA, 126 Fuzzy DEMATEL, 110 Fuzzy Integral, 140 Fuzzy Logic, 72, 112, 122, 153, 245, 367 Fuzzy Mathematical Programming, 58, 60, 310, 332 Fuzzy MCDM, 177 Fuzzy Measures , 280 Fuzzy Numbers, 61, 194 Fuzzy Reasoning, 108 Fuzzy Set Extension Of DRSA, 61 Fuzzy Sets, 64, 118, 127, 138, 143, 193, 195, 197 Fuzzy VIKOR, 197 Gaussian Process Modeling, 203 Generating Alternatives, 109 Genetic Algorithms, 62, 74, 109, 153, 170, 238, 257, 267 GIS, 225, 322 GIS-MCDA Integration, 225 Global Optimization, 292 Goal Programming, 64, 90, 118, 143, 157, 195, 200–202, 210, 328, 336 Governance, 346 Graph Theory, 325 GRASP Algorithm, 76 Greenhousegas Emissions, 119 Grey Numbers, 205 Group Decision Making, 46, 72, 220 Group Decisions Making, 64, 110, 118, 127, 273, 300, 304, 372 guided search, 99, 168 Facility Location Problem, 327 Factory Layout, 62 FAHP, 104, 221, 319 Favelas, 221 Feasibility Approach, 213 Feature Selection, 187 Federal Funds Distribution, 165 Finance, 176, 290, 349 Financial Performance, 73 Fleet Availability, 233 Flight And Maintenance Planning, 233 Food Supply Chains, 239 Forest Management, 69, 206, 273 H¨older Metric, 224 Harvest Scheduling, 69, 210 Healthcare, 181, 227 Heuristic Search, 174, 274 Hierarchical Criteria, 334 Hierarchical Optimization, 111 Higher Education Institution, 338 Historical Industrial Buildings, 345 Hospital Site Selection, 112 Human Development Index, 276 Human Resources, 123 Humanitarian Logistics, 78, 138, 327 380 Hybrid Approaches, 65, 66, 81, 259, 287 Hybrid Eigenvalue, 354 Hybrid Framework, 209 Hydroelectric Plant, 319 Hyper Radial Visualisation, 364 Hyperspace Pareto Frontier, 364 Learning Utility Functions, 361 Life Cycle Analysis, 240 Life Testing, 47 Line Balancing Control, 248 Linear Approximation, 307 Linear Model, 183 Linear Programming, 89, 92, 98, 123, 191, 222, 227, 288, 306, 324 Linear Scalarization, 137 Linear Transformation, 283 Linguistic Variables, 194 Location, 87, 180, 223, 348 Logistics, 116, 202 Loss Aversion, 134 LR-Type Fuzzy Variables, 196 Ideal Proportions, 357 Identifying Objectives, 151, 304 Immigration, 153 Immune Algorithm, 154 Impacts Of Statutory Planning, 71 Importance Weights, 110 Imprecise Information, 82, 92 Improving Method, 188 Incomplete Information, 220 Inconsistency, 321 Indifference Curve, 357 Indifference Thresholds, 129 Innovation Strategy, 177 Integer Linear Programming, 84 Integer Programming, 106, 114, 189, 294, 376 Intellectual Capital, 156 Intelligent Decision System, 121 Intensity Modulated Radiation Therapy, 244 Interaction Among Criteria, 219, 280, 358 Interactions, 370 Interactive EMO, 209, 361 Interactive Methods, 102, 106, 113–115, 117, 155, 279, 282, 307, 331, 348 Interactive Multi-Objective Optimization, 277 Interrelations, 342 Interval-Valued Functions, 205 Invasive Alien Species, 225 Inventory Routing Problem, 116 Inverse Optimization, 363 Investment Analysis, 176 Irrigated Olive Grove, 120 MACBETH, 260, 370 Machinery Industry, 96 Macroscopic Model, 261 Man-Machine Algorithm, 357 Management Accounting And Auditing, 201 Map Comparison, 173 Marine Engineering Systems, 122 Marketing, 234 Markov chains, 265 Matching, 88 Mathematical programming, 54, 97, 125, 187, 265 Matrix Representation, 179 Matroids, 85 Measurement, 107, 148, 150 Measures Of Sensitivity And Efficiency, 329 Mechanical Engineering, 192 Meta-Goal, 118, 143, 195 Metaheuristics, 100, 111, 222, 268 Microsoft Excel, 230 Minimal Size, 285 Mixed Integer Multi-Objective Programming, 233 Mixed Integer Programming, 49, 106, 223, 227, 232, 242, 293, 294 Model Testing And Validation, 250 Model-Based Design Of Experiments, 249 Monte Carlo Simulation, 82, 153 Morphological Analysis, 194 Multi-Actor Multi-Criteria Analysis, 91, 350 K-Min Objectives, 259 KLD Rating, 73 Knapsack, 206 Lagrangian Duality, 164 Land Use Planning, 206, 267 Large Scale Optimization, 52 381 Multi-Agent Systems, 102, 130 Multi-Attribute, 236 Multi-Attribute Utility Theory, 67, 131, 146, 149, 161, 186, 208, 237, 340, 368 Multi-Attribute Value Theory, 51, 72, 234, 288 Multi-Choice Goal Programming, 129 Multi-Constraint Levels, 179 Multi-Criteria Combinatorial Optimization, 190, 362 Multi-Criteria Decision Aid, 77, 165, 178, 219, 280, 321, 322, 327, 336 Multi-Criteria Decision Analysis, 57, 72, 84, 86, 98, 128, 165, 182, 204, 205, 216, 217, 226, 237, 240, 245, 255, 275, 276, 301, 315, 342, 353, 375 Multi-Criteria Decision Making, 53, 55, 58, 59, 65, 68, 82, 103, 106, 110, 127, 131, 142, 145, 155, 160, 162, 181, 192, 209, 230, 231, 235, 242, 258, 271, 273, 284, 298, 314, 334, 337, 347, 353, 354, 363, 366, 367 Multi-Criteria Decision Model, 79, 285 Multi-Criteria Evaluation, 65, 343 Multi-Criteria Evaluation Model, 260 Multi-Criteria Hierarchy Process, 272 Multi-Criteria Investment Problem, 224 Multi-Criteria Logic, 329 Multi-Criteria Mapping, 377 Multi-Criteria Model Predictive Control, 308 Multi-Criteria Ranking, 183 Multi-Criteria Sorting, 123, 222, 246 Multi-Dimensional, 107 Multi-Dimensional Knapsack Problem, 259, 344 Multi-Methodology, 165 Multi-Objective Optimization, 357 Multi-Objective Shortest Path Problem, 135 Multi-Objective Assignment, 152 Multi-Objective Combinatorial Optimization, 152 Multi-Objective Differential Evolution, 213 Multi-Objective Genetic Algorithms, 184 Multi-Objective Knapsack, 152 Multi-Objective Linear Programming, 179 382 Multi-Objective Mixed Integer Programming, 365 Multi-Objective Optimization, 50, 54, 56, 60, 62, 69, 74, 78, 84, 88, 89, 98, 100, 105, 106, 109, 111, 114, 116, 117, 124, 130, 132, 135, 147, 153, 154, 158, 164, 166, 167, 169, 175, 179, 189–192, 196, 199, 203, 207, 211, 214, 231, 244, 247–249, 251, 252, 255–257, 259, 261, 264, 266, 268–270, 273, 279, 282, 286, 287, 289–292, 294, 307, 320, 323, 324, 326, 328, 331, 333, 339, 344, 355, 363, 373, 374, 376 Multi-Objective Shortest Path Problem, 174, 214 Multi-Parameter Functions, 129 Multi-Project Scheduling, 235 Multi-Purpose Water Resource, 104 Multi-Scenario, 136 Multi-Stakeholder Decision Making, 304 Multi-Variate Decision Analysis, 186 Multimoora, 290 Multiplicative Analytic Hierarchy Process, 364 Multiplicative Form, 161, 290 Mutation, 209 Nadir Point, 147, 169, 189 National Transport Planning, 346 Natural Gas Pipeline, 275 NAUTILUS Method, 209 Nearly Subconvexlikeness, 137 Negotiation, 47 Net Present Value, 356 Network Ration Map, 352 Network Resilience, 326 Neural Networks, 108 New Product Development, 177, 250 NIMBUS, 102, 279 Non-Dominated Frontier, 184, 233 Non-Dominated Set, 211 Non-Dominated Surfaces, 363 Non-Linear Programming, 119 Non-Linear Multi-Objective Optimization, 56, 188, 281, 287, 307 Non-Linear Scalarization, 80 Non-Regular Problems, 171, 207 NSGA-II, 209 Objectives, 198 Obsolescence, 128 Offshore Wind Farm, 202 Open And Distance Learning, 238 Operator Allocation, 248 Optimal Control, 251, 308 Optimality Conditions, 171 Optimization, 325 Ordered Weighted Av, 88 Ordinal Information, 284 Ordinal Regression, 178 Organizing Objectives, 151 Outer Approximation, 333 Outranking Methods, 68, 160, 163, 228, 272, 334, 337, 375 Outranking Relation, 278 Over-Fitting, 250 OWA Operator, 135 Point Estimates, 249 Polynomial Model, 183 Portfolio Optimization, 231, 298 Portfolio Selection, 55, 58, 184, 196, 363 Positional Voting Systems, 313 Positively Homogeneous Functions, 244 Possibilistic Framework, 252 Possibilistic Model, 310 Possibility Theory, 252 Power Generation, 232 Preference Aggregation, 118, 131, 226, 317 Preference Disaggregation, 79, 245, 278 Preference Information, 169, 201, 277, 308 Preference Learning, 102, 219, 222, 234, 245, 302, 366, 369 Preference Modeling, 157, 219, 278, 308, 336 Preference-Based Evolutionary Algorithms, 166, 291 Preferences, 87, 173, 200, 207, 215 Preferences And Utilities, 329 Premature Newborns Care, 368 Preservation Region, 305 Preventive Maintenance, 74 Price Volatility, 212 Pricing, 301 Principal Component Analysis, 307 Principle Of Tolerance, 350 Prioritisation Methods, 266 Prioritizing Objectives, 151, 221 Privacy, 243 Problem Structuring, 131, 148, 150, 260, 341 Product Development, 128 Production, 143, 248, 330 Production Cost, 232 Production Scheduling, 50, 74 Project Portfolio Selection, 60 Project Risk, 79 Project Scheduling, 235 PROMETHEE, 91, 163, 265, 312, 358, 371 Proper Epsilon-Efficiency, 137 Prospect Theory, 371 Provable Approximation Guarantees, 316 Proximity-Based Decision Rules, 308 Pseudo-Boolean Search Space, 361 Psychology, 107 Pairwise Comparison, 157, 172, 230, 321 ParadisEO-MOEO, 252 Parallel Algorithm, 281 Parallel Computing, 175 Parallel Multi-Objective Optimization, 295 Parallelism, 100 Parameter Precision, 249 Parameter-Dependent Problem, 360 Parametrization, 348 Pareto Dominance , 252 Pareto Equilibria, 296 Pareto Local Search Algorithm, 252 Pareto Optimal Investment Portfolio, 224 Pareto Optimal Solutions, 78, 125, 166, 253, 279, 323, 328, 355 Pareto Set, 255, 285 Partial Preorder, 334 Participatory Evaluation, 241 Participatory Multi-Criteria Evaluation, 288, 377 Path-Following Methods, 287, 360 Penalty And Reward Functions, 318 Penalty Methods, 170 Performance Measurement, 86, 175, 229, 353 Performance Targets, 318 PISA Survey, 228 Plant Capacity Planning, 119 Plaster Waste, 297 Plastic, 325 383 Public Policy, 77, 346 Public Safety, 68 Public Transport, 377 Robust Ordinal Regression, 178, 219, 220, 278, 322, 340, 358 Robust Supply Chains, 231, 327 Robustness Analysis, 278 Robustness And Sensitivity Analysis, 231, 324, 335, 340, 341 Rough Sets, 140 Routing, 75, 76 Rule Extraction, 218 Rule-Based Knowledge Acquisition, 218 Runtime Analysis, 361 Rural Planning, 162 Qualitative Multi-Criteria Model, 183, 317 Qualitative Reasoning, 127 Quality Indicators, 188 Quickest Flow, 261 Radial Projection, 305 Rank Reversal, 312 Ratio Analysis, 290 Ratio Of Directional Derivatives, 357 Reachability, 62 Real Options Analysis, 262 Recycling Facilities, 53 Redistribution Mechanisms, 130 Reference Hyperplane, 305 Reference Multifunction Method, 308 Reference Point, 99, 168, 169, 267, 276, 289–291, 316, 321, 349, 371 Reference Point based NSGA-II, 364 Regions, 354 Reinforcement Learning, 255 Relational Modeling, 317 Relative Importance, 357 Renewable Energy, 232 Replacement Cost, 356 Representation Quality, 211 Representative Set, 191, 211 Resampling, 168 Resort Development Investment, 356 Restoration, 67 Retailing, 377 Returns To Sclae, 305 Revealed Preferences, 212 Revenue Management, 301 Revenue-Risk Model, 329 Reverse Auctions, 236 Risk, 216, 275 Risk Analysis, 101, 141, 176, 263, 301, 319 Risk Assessment, 70 Risk Communication, 372 Risk Management, 241, 243, 370, 372 Road Safety, 158 Robust Elicitation, 321 Robust Optimization, 80 Safest Flow, 261 Satisfying Philosophy, 336 Savage’s Criteria, 224 Scalarization, 54, 56, 144 Scale Compatibility, 134 Scenario Planning, 194 Scenarios, 194, 230–232, 241, 279, 327 Scheduling, 54, 63, 144, 271 Scientific Uncertainty, 372 Scoring Rules, 313 Screening, 73 Semidefinite Programming, 89 Sensitivity Analysis, 53, 78, 94, 125, 146, 232, 237, 288, 305 Sensory Attributes Prediction, 250 Sequential Solution Approach, 213 Service Quality, 140 Set Covering Problem, 49 Shortest Path, 75, 175, 214, 247, 335 Simple Elitist Evolutionary Algorithm, 252 Simulated Annealing, 268 Simulation-Based Optimization, 99, 168 Simultaneous Solution Approach, 213 Singular Value Decomposition, 307 SIR, 350 Situational Judgment, 107 SMAA, 92, 340, 358 Smart Devices, 177, 232 Smart Home System, 352 Smart-Swaps, 134 Social Decision Making, 299 Socially Responsible Investment, 55, 126, 299, 349, 363 Soft-Computing, 184, 194 Software, 98, 230, 314 Software Protocol, 295 384 Solution Concepts, 136 Solution Diversity, 109 Southern Spain, 120 Spanning Tree, 124 Spatial Multi-Criteria Evaluation, 131, 206, 323, 345, 375 Stability Radius Of Portfolio, 224 State Of The Art, 202 Statistics, 132 Stochastic, 356 Stochastic Dominance Constraints, 320 Stochastic Knapsack, 351 Stochastic Linear Program, 339 Stochastic Multi-Objective Acceptability, 178 Stochastic Programming, 80, 320, 331 Stochastic Systems, 168 Stopping Rule, 188 Strategic Decisions, 341 Strategic Game, 296 Strategic Information Systems, 93 Strategic Planning, 77, 341 Structural Modeling, 268, 342 Structuring Decisions, 108, 151, 297 Substitutability between components, 276 Supervised Learning, 218 Supply Chain Management, 66, 93, 143, 193, 301 Supported Non-Dominated Solution, 211 Surrogate Modeling, 81, 132, 203 Surrogate Relaxation, 344 Sustainability, 59, 77, 93, 204, 239, 306, 330, 346, 353, 371 Sustainability Assessment, 104, 158, 280, 345, 353, 375 Sustainable Supply Chains, 59, 347 Synhronous Methods, 348 Synthetic indicators, 276, 299, 349 System Integration, 232 Systemic Decision Making, 350 Tabu Search, 95 Task Allocation, 257 Third Party Logistic Supplier, 303 Time Dependent Data, 261 Time Surplus, 133 Timetable Information, 247 Timetabling, 90 Tolerance Distribution, 350 385 TOPSIS, 86, 197, 262, 263 Town Regeneration, 377 Trade-Offs, 232 Traffic, 154 Train Networks, 247 Transaction Costs, 320 Transport Policy, 346 Transportation, 75, 116 Transshipment, 261 Travelling Salesman, 325 Triangular Fuzzy Numbers, 252 Triple Bottomline, 364 Two Types Of Error, 179 Two-Phase Method, 365 Uncertain Pareto Dominance, 203 Uncertainty, 53, 80, 252 Uncertainty Handling, 279, 288 Uncertainty Modelling, 193, 196, 208, 216, 237 Uranium Processing, 67 User Equilibrium, 133 UTADIS, 123, 246, 340 Utility, 185 Utility Theory, 101, 159, 215, 216 Value Focus Thinking, 131, 148, 150, 185, 198, 297 Value Function, 115, 373 Variable Reduction, 283 Variational Problems, 270 Vector Optimization, 137 Vehicle Routing Problem, 95, 252 Vehicular Ad Hoc Networks (VANET), 295 VIKOR, 112, 229, 367 Visegrad Four, 354 Visual Steering, 364 Vulnerability, 241 Waste Water Water Water Collection Problem, 76 Infrastructure Planning, 215, 237 Resources, 318 Resources Management, 149, 197, 237, 323, 326, 372 Web Architecture, 376 Weighted product, 86 Weighted Sum, 56, 146 Weights, 51, 215 Wellbeing, 61 WINGS, 342 Workforce Scheduling, 200 386 List of Authors ¨ orni, Anssi O¨ MO3E, 51 ¨ Ozpeynirci, Selin FR3E, 246 C ¸ o¨l, G¨ ul¸cin MO2E, 200 Altuzarra, Alfredo MO5H, 350 Alves, Maria Jo˜ao TU3F, 111 TU3G, 117 Amer, Noha TH4F, 257 Andersen, Kim Allan TH2B, 351 Andina, Diego FR3G, 160, 162, 163 TH3G, 145, 337 Andr´es Pay´an, Ana MO5H, 53 Angilella, Silvia TU2E, 280 TU4D, 92 Angulo-Guerrero, Mar´ıa Jos´e TH4C, 276 Anic, Ivan WE2B, 228 Ant´on, Jos´e FR3G, 162 TH3G, 337 FR3G, 160, 163 TH3G, 145 Antunes, Paula TU2B, 377 Aouni, Belaid MO2E, 336 MO2B, 201 Arenas-Parra, Mar TH2D, 64 TU2D, 195 Arriaza, Manuel TH4B, 120 Autran M. Gomes, Luiz F. FR3E, 123 WE1E, 306 Avinadav, Tal TU5D, 301 Aydinbeyli, Yakup Emre TH4F, 63 Azarm, Shapour TU4D, 269 Abdelkader, Mendas MO2D, 226 AbuDahab, Khalil MO5H, 218 Afsordegan, Arayeh TU5B, 127 Agell, N´ uria TH3C, 245 TU5B, 127 Aguar´on, Juan MO5H, 300 Aguayo, Ernesto TH3D, 82 Akyol, Emine TU5E, 144 Al¸cada-Almeida, Lu´ıs TH4E, 375 Alba, Enrique MO5H, 166 WE2E, 295 Alcal´a, Juli´an TU5E, 268 Alegr´ıa Mej´ıa, Lol-chen MO4F, 265 MO4F, 240 Alencar, Marcelo WE2D, 275 WE2D, 297 Allmendinger, Richard TH2E, 207 TU4F, 130 Almeida, Adiel WE2D, 275 Almeida-Filho, Adiel WE2D, 74 Almulhim, Tarifa WE2C, 110 Alonso, Pedro WE1B, 338 387 B¨aker, Bernard TH3F, 343 B¨ using, Christina TU3E, 316 Backovic, Marko WE2B, 228 Baeza-Sampere, Ismael TU4G, 126 WE1C, 310 Bahri, Oumayma FR3D, 252 Bai, C. M. TU5D, 66 Balaguer, Rosario TU4G, 55 Bamford, David TH4C, 353 Bana e Costa, Carlos TH1A, 309 TU4E, 370 Banasik, Aleksander TU3B, 239 TH4D, 227 Bandet, Jean TU4C, 71 Barfod, Michael Bruhn TU5B, 141 Batista dos Santos, Lucelina MO4C, 374 Baucells, Manel TU4E, 159 Bayazit, Ozden TU3B, 303 Bazgan, Cristina FR2E, 285 Beck, Joakim TH3E, 132 Belmokaddem, Mustapha MO2E, 129 Belton, Valerie TU5C, 139 Ben Abdelaziz, Fouad TH4F, 257 Ben Amor, Nahla FR3D, 252 Berbel, Julio MO5H, 212 Berm´ udez, Jos´e D. WE1F, 196 Bernal, Astrid FR2D, 368 Bertomeu Garc´ıa, Mercedes MO3D, 210 Bertsch, Valentin WE1D, 288 WE1D, 241 Bilbao, Amelia TH2D, 64 TU2D, 195 Blanco, Victor WE1G, 89 Boggia, Antonio FR2C, 322 Bohanec, Marko FR2G, 183 TU2B, 317 Borrion, Herv´e MO4G, 243 Bottero, Marta TU2E, 280 TU3B, 345 Bouslah, Kais MO2G, 73 Branke, Juergen TU3F, 369 Brauers, Willem K. M. MO2B, 290 Bravo Sell´es, Milagros TU4G, 299 Brison, Valerie MO4F, 173 Brockhoff, Dimo MO3F, 361 Brzostowski, Jakub TH3D, 359 Buitrago, Eva WE2G, 296 Buruk, Yeliz TU3G, 54 Buxad´e, Carlos TU4C, 146 C´ordoba, Mar´ıa-Adelaida FR2D, 368 Caballero Fern´andez, Rafael FR2F, 152 FR3C, 60 388 MO3D, 69 TU3C, 78 Cabello, Jose Manuel MO2G, 349 MO3F, 291 Cabrera G., Guillermo TU5E, 244 Cadenas, Jos´e Manuel WE1F, 184 Calvete, Herminia I. TU3F, 83 Calvo, Clara WE1F, 58 Camanho, Roberto WE1E, 306 Camanho, Thomaz M. WE1E, 306 ´ Camino-Saco, Angeles WE1B, 172 Campoy-Mu˜ noz, Pilar TU2F, 153 Cano, Javier MO4G, 101 TU2F, 323 Caraballo, M. Angeles WE2G, 296 Casas P´aez, Pedro Nicol´as WE1B, 128 Castillo, Mario FR2D, 368 Cavalcante, Cristiano WE2D, 74 Cekyay, Bora TH4B, 96 Cerqueus, Audrey MO4D, 344 Chaabane, Djamal FR3C, 84 FR3D, 339 Chandra, Arjun TU4F, 130 Chang, Chiao-Chen TU4B, 356 Charpak, Nathalie FR2D, 368 Chen, Chiau-Ching TH4B, 94 Chen, Yu-Wang MO4E, 121 MO4B, 250 MO5H, 218 WE2G, 108 Chen, Yue MO4E, 121 Chern , Yuching MO5H, 367 Chernonog, Tatyana TU5D, 301 Cheung, Alan K. L. MO2D, 75 Chiazzaro, Mauricio TU2F, 323 Chikumbo, Oliver MO4E, 364 Chin, Yang-Chieh TU4B, 140 Chmielewski, Hana FR2F, 109, 326 Chung, Eun-Sung FR3F, 197 Church, Richard TH2B, 214 Ciffroy, Philippe WE2C, 72 Cimen, Emre TH3C, 187 Cipriano Rodrigues, Teresa MO3C, 260 Cisneros, Jos´e Manuel TH3G, 145, 337 Cl´ımaco, Jo˜ao WE1E, 204 Claassen, G.D.H. (Frits) TU3D, 242 TH4D, 227 TU3B, 239 Clemente, Th´arcylla WE2D, 275 Coelho Silva, Ricardo TH2D, 332 Coenen, Thomas TU2G, 251 Coll-Serrano, Vicente TU4G, 126 WE1C, 310 Colombo, Federico 389 FR3G, 163 Colorni, Alberto TH3D, 302 Comes, Tina TU5D, 327 WE1D, 241 Corrente, Salvatore TH3B, 272 TU4D, 92 WE2B, 358 FR2C, 322 TU2E, 280 TU3F, 369 Cortes Aldana, Felix Antonio WE1B, 128 Cremades, L´azaro V. TU5B, 127 Critto, Andrea MO5H, 70 WE2C, 72 Cruz Corona, Carlos TH2D, 332 TU2D, 194 Dehnokhalaji, Akram WE1C, 305 Del Vasto, Luis WE2B, 334 Dellnitz, Michael TU2G, 360 Dias, Luis FR2D, 234 TU2E, 220 TH4E, 375 Diaz-Madro˜ nero, Manuel TH2D, 143 WE1F, 193 Dietermann, Ansgar TH3F, 343 Dirks, Kim N. MO2D, 75 Dopazo, Esther TH2D, 118 Dosal Vi˜ nas, Elena MO5H, 53 Drugan, Madalina M. MO2F, 255 Duarte, Abraham TU3D, 76 Dudas, Catarina WE1D, 330 Duh, Kevin MO5H, 254 Durillo, Juan J. MO5E, 100 D¨achert, Kerstin TH2G, 105 D´ıaz-Balteiro, Luis MO3D, 210 da Silva Vieira, Ricardo TU2B, 377 Dall’O, Giuliano TH3G, 77 Dandurand, Brian TH2E, 164 De Armas, J´esica TU2D, 256 de Castro Pardo, M´onica MO2D, 304 De los R´ıos, Lisandro FR3G, 163 De Schepper, Ellen TU5C, 293 De Smet, Yves WE2B, 312 TH4C, 158 Deb, Kalyanmoy MO2F, 99, 168 MO4E, 364 WE1D, 330 Ehrgott, Matthias TH2G, 191 MO2D, 75 TH2B, 351 TH4G, 48, 133 TU5D, 231 TU5E, 244 TU5F, 333 El badraoui, Khalid MO2G, 73 El-Ghazali, Talbi FR3D, 252 El-Haj Ben-Ali, Safae WE1G, 89 Emelichev, Vladimir MO2G, 224 Erb, Stephan 390 TU5G, 175 Ergu, Daji FR2B, 142 Escobar, Mar´ıa Teresa MO5H, 300 Escobar-Toledo, Carlos Enrique MO4F, 240 MO4F, 265 Eskova, Olga MO4G, 329 TU2C, 357 MO3D, 69 G´omez-Garc´ıa, Esteban WE1B, 172 G´omez-Lim´on, Jos´e A. TH4B, 120, 156 Gal´e, Carmen TU3F, 83 Galand, Lucie WE2F, 135 Galbiati, Lorenzo TU2F, 323 Gallina, Valentina MO5H, 70 Gandibleux, Xavier FR2E, 365 MO4D, 344 TH4G, 52 TU5C, 293 Garc´ıa Bernabeu, Ana M. TU4G, 55, 299 Garc´ıa P´erez, Mar´ıa Dolores WE2G, 223 Garc´ıa-Alonso, Carlos R. TU2F, 153 Garc´ıa-Segura, Tatiana TU5E, 268 Gargallo, Pilar MO5H, 350 Garnevska, Elena FR2D, 186 Garrido, Maria del Carmen WE1F, 184 Gavranis, Andreas WE2E, 233 Gayt´an-Iniestra, Juan TU3C, 78 Geiger, Martin Josef FR3C, 116 Geldermann, Jutta TH4E, 371 Genova, Krasimira TH4D, 376 Gerdessen, Johanna TH4D, 227 TU3D, 242 Ghaderi, Mohammad TH3C, 245 Gila-Arrondo, Ar´anzazu Farion, Ken TH2C, 181 Faustino, Ana FR3F, 318 Fern´andez Hern´andez, Pascual WE2G, 180, 223 Fern´andez-Hern´andez, Jos´e MO2F, 188 TU3E, 281 Ferreira, Rodrigo WE2D, 74 Ferretti, Valentina TU3B, 345 TU4C, 131 TU2E, 280 Fiala, Petr WE1G, 167 Fichtner, Wolf WE1D, 288 Figueira, Jose TH4D, 259 WE2B, 358 Filipiˇc, Bogdan TU3F, 203 Fraga, Eric TH3E, 132 Fuentes Rojas, Ever Angel MO5H, 325 Fulga, Cristinca TU4G, 298 G¨ ul, Sait WE1B, 112 G´omez, Carlos Mario MO5H, 212 G´omez, Trinidad FR3C, 60 391 TU3E, 281 Gim´enez, Juan Carlos MO3D, 210 Ginestar, Concepcion MO2D, 304 Gocuklu, Gulcan TU3G, 54 Goetzmann, Kai-Simon TU3E, 316 Goienetxea, Ainhoa WE1D, 330 Goldsztejn, Alexandre TH2F, 287 Gon¸calves, Marcos E.L. FR3E, 123 Gonz´alez-Pach´on, Jacinto TU2B, 157 Goodman, Erik MO4E, 364 Gorobetz, Mikhail MO5H, 154 Gort´azar, Francisco TU3D, 76 Gouliashki, Vassil TH4D, 376 Gourv`es, Laurent MO4D, 85 Granvilliers, Laurent TH2F, 287 Grau, Juan FR3G, 160, 163 TH3G, 145 FR3G, 162 TH3G, 337 Grebe, Tabea TH2E, 307 Greco, Salvatore FR2C, 322 TU2E, 178 TH3B, 219, 272 TU2E, 280 TU3E, 277 TU3F, 369 TU4D, 92 WE2B, 358 Guerrero-Baena, Mar´ıa Dolores TH4B, 156 Gurevsky, Evgeny TH4G, 52 Gurgel, Andr´e WE2D, 68 Guti´errez, C´esar MO4C, 137 Gutierrez-Martin, Carlos MO5H, 212 Gutjahr, Walter FR3D, 320 H¨am¨al¨ainen, Raimo MO3E, 134 Habenicht, Walter MO2C, 106 Hakanen, Jussi TU4F, 81 Handl, Julia TH2E, 207 TH3C, 217 TU4F, 315 Hartikainen, Markus MO4B, 282 Hasannasab , Maryam TU5F, 333 Hasg¨ ul, Servet MO2E, 200 Hazin, Luciana WE2D, 297 Henggeler Antunes, Carlos TU3G, 117 Henig, Mordecai TU5D, 301 Hern´andez Jim´enez, Mar´ıa Beatriz MO4C, 171 Hern´andez, Elvira MO4C, 125 Hern´andez-D´ıaz, Alfredo G. TU3D, 76 FR2F, 152 Hernandez, Monica MO3D, 69 ´ Hinojosa, Miguel Angel TU3D, 76 Hirschberger, Markus MO2G, 363 Ho, Phong TH4F, 62 Horenkamp, Christian TU2G, 360 392 Hu, T. L. TU5D, 66 Hu, Weiwei TU4D, 269 Huang, Chi-Yo TU4B, 177 Huang, Chung-An TU3C, 352 Huber, Sandra FR3C, 116 Huerga , Lidia MO4C, 137 Huy, Truong TH4F, 62 Jim´enez, Bienvenido MO4C, 137 John, Agah TU4F, 315 Jones, Dylan TU1A, 202 Juutinen, Artti TH2F, 206 K¨obis, Elisabeth TH2F, 80 Kabak, Ozgur TH4B, 96 Kadzinski, Milosz MO5E, 278 TH3B, 340 Kahraman, Y.Rıza MO3B, 57 Kaliszewski, Ignacy MO3F, 192 Kamisli Ozturk, Zehra MO3B, 238 Kandakoglu, Ahmet MO3B, 57 TU5E, 86 Kandakoglu, Makbule TU5E, 86 Kanellopoulos, Argyris TU3B, 239 Kao, Yu-Sheng TU4B, 177 Karakaya, G¨ ul¸sah TU3G, 155 Karelkina, Volha MO2B, 348 Karlsson, Alexander MO2F, 168 Karpak, Birsen TU3B, 303 TU5B, 59 WE1E, 347 TU3B, 93 Kasimbeyli, Refail MO4B, 56 TH2B, 97 Keeney, Ralph MO1A, 151 Keshvari, Abolfazl WE1C, 283 Ichiro, Nishizaki TU4E, 258 MO3C, 161 Ide, Jonas TU5D, 231 Iranzo, Jos´e A. TU3F, 83 Isern, David TH4E, 366 Isigonis, Panagiotis MO5H, 70 WE2C, 72 Ismaili, Anisse MO4D, 124 WE2F, 135 Ivorra, Carlos WE1F, 58 Jablonsky, Josef TH2C, 314 Jamain, Florian FR2E, 285 Javier, Fiallos TH2C, 181 Jensen, Anders Vestergaard TH4C, 346 Jermann, Christophe TH2F, 287 Jim´enez L´opez, Mariano TH2D, 64 TU2D, 195 Jim´enez, Antonio TH3D, 82 TU5C, 67 393 Keyhani, Mohammad Hossein WE2F, 247 Kiczkowiak, Tomasz MO3F, 192 Kirilov, Leoneed TH4D, 376 Kirlik, Gokhan TU5F, 147 Kiszova, Zuzana WE2C, 354 Klamroth, Kathrin TH2F, 80 MO4B, 282 TH2G, 105, 362 TH4D, 259 Knowles, Joshua TH2E, 207 TU4F, 315 Ko, Yu-Chien WE2C, 61 Ko¸canli, M¨ uzeyyen Melek TH4F, 63 Kokko, Tommi TU4F, 81 Koksalan, Murat MO2C, 114 FR3C, 189 FR3E, 246 MO2C, 115 MO4B, 113 TU3G, 155 Korhonen, Pekka WE1C, 283 MO2C, 114 MO3E, 51 Korotkov, Vladimir MO2G, 224 Kou, Gang FR2B, 45, 142 Kozanidis, George WE2E, 233 Krampf, Peter MO3G, 185 Krieter, Joachim TU4C, 146 Krupi´ nska, Katarzyna FR2G, 87 Krym, Eduardo WE2D, 275 Krzeszowska, Bogumila TU4D, 235 Kuhn, Kenneth WE2F, 335 Kunz, Reinhard MO3G, 148 MO3G, 150 Kuo, R. J. TU5D, 66 L´opez Agudo, Luis Alejandro MO2B, 289 L´opez Montoya, Rub´en MO4C, 125 L´opez-Redondo, Juana MO2F, 188 TU3E, 281 L´opez-S´anchez, Ana Dolores TU3D, 76 Laguna, Manuel FR2F, 152 Lahtinen, Tuomas MO3E, 134 Lamata, Mar´ıa T. TU2D, 194 Lami, Isabella TU2E, 280 Lancinskas, Algirdas WE2G, 180 Langhans, Simone MO2D, 149 TU5C, 208 Laryea, Rueben FR3E, 79 Lauven, Lars-Peter WE1D, 119 Lavygina, Anna TH2E, 207 Le Sage, Tanya MO4G, 243 Lee, Boram FR3F, 197 Leleur, Steen TH4C, 346 Leon, M. Amparo MO3D, 69 Lerche, Nils TH4E, 371 394 Levchenkov, Anatoly TH4F, 271 Lewis, Peter TU4F, 130 Lienert, Judit MO2D, 149 FR3F, 237 TU3G, 215 TU5C, 208 Liern, Vicente FR3C, 60 WE1F, 58 Lilti, Jean-Jacques MO2G, 73 Lin, Changsheng FR2B, 45 Lin, T. Tyrone TU4B, 356 Lin, Yi-Chun TH4B, 94 Liu, Jinyan FR2C, 321 Liu, Shuang FR3E, 372 Lizin, Sebastien TU5C, 293 Llamazares, Bonifacio TH2C, 313 Logist, Filip TU2G, 251 Lokman, Banu FR3C, 189 MO2C, 114, 115 Long, Pham MO3C, 236 Lopes, Diana F. TU4E, 370 Lozano, Carlos A. FR3B, 263 Lu, Ming-Tsang TU4B, 65 Luque, Mariano MO2B, 289 MO3F, 291, 355 MO5H, 166 TH4C, 276 TU3E, 169 Lust, Thibaut FR2E, 199 M’Zali, Bouchra MO2G, 73 TU4G, 126 M¨onkk¨onen, Mikko TH2F, 206 M´armol, Amparo Mar´ıa MO5H, 136 WE2G, 296 M´endez Rodr´ıguez, Paz MO2G, 349 TU4G, 55, 126 Maag, Volker TH2E, 307 Macedo, Marcelo G. C. WE1E, 306 Macharis, Cathy MO4F, 91 Machuca, Enrique WE2F, 274 Maheshwari, Vaibhav TH3E, 249 TU5F, 213 Malecki, Jozef TH3E, 122 Mandow, Lorenzo TU5G, 328 WE2F, 274 Manel, Maamar WE2E, 264 Manotas , Diego F. FR3B, 263 FR3B, 262 Manotas , Luis A. FR3B, 262 ´ Marcenaro-Guti´errez, Oscar MO2B, 289 Marcomini, Antonio MO5H, 70 WE2C, 72 Marin, Lucas TH4E, 366 Marins, Fernando WE1B, 338 Maroto, Concepci´on MO2D, 304 MO3D, 273 Marqu´es, Inmaculada 395 MO3D, 311 Marr, Norman FR2D, 186 Mart´ın Fern´andez, Paula TU4C, 146 Mart´ın, Jacinto MO5H, 47 Mart´ın, Miguel C. TU5C, 67 Mart´ınez Gamboa, Jeyson Andres MO5H, 325 Mart´ınez-Ortigosa, Pilar MO2F, 188 TU3E, 281 Martin, Benjamin TH2F, 287 TU5C, 293 Martin, Rodrigo FR3G, 160 Martinez, Raquel WE1F, 184 Martins, Luiz Geraldo Biagioni WE1E, 306 Masakazu, Ohmi TU4E, 258 Masegosa, Antonio D. TU2D, 194 Mason, Andrew TU5E, 244 Massei, Gianluca FR2C, 322 WE1E, 225 Matarazzo, Benedetto TU3E, 277 Mateo, Manuel TH4G, 50 Mateos, Alfonso TH3D, 82 TU5C, 67 Matuschke, Jannik TU3E, 316 Mazurek, Jiˇr´ı TU3C, 284 WE2C, 354 Mazziotta, Adriano TH2F, 206 McGillis, Sheila MO2B, 201 Mebrek, Fatma FR3D, 339 Medrano, F. Antonio TH2B, 214 Mej´ıa-Argueta, Christopher TU3C, 78 Meli´an, Bel´en TU2D, 256 Melloul, Sakina MO2E, 129 Merel, Aur´elien TH4G, 52 Michalowski, Wojtek TH2C, 181 Michnik, Jerzy FR2B, 342 Miettinen, Kaisa MO3E, 279 MO4B, 282 TH2E, 102 TH2F, 206 TU2G, 253 TU3E, 169 Mikhailov, Ludmil MO3B, 266 WE2C, 110 Milan, Lauriane MO4F, 91 Mileva-Boshkoska, Biljana FR2G, 183 Miroforidis, Janusz MO3F, 192 Mitchener-Nissen, Timothy MO4G, 243 Mitsuhiro, Imai MO3C, 161 Mlakar, Miha TU3F, 203 Molina Luque, Juli´an FR2F, 152 MO3D, 69 TU3C, 78 Molines, Nathalie TU4C, 71 Mondini, Giulio TU3B, 345 Monnot, J´erˆome MO4D, 85 396 Monroy, Luisa MO5H, 136 WE2G, 296 Montero, Javier FR3G, 138 Montibeller, Gilberto TU2C, 341 TU4C, 131 Moradi, Siamak TH4G, 48 Moreno P´erez, Jos´e A. TU2D, 256 Moreno, Antonio TH4E, 366 Moreno-Jim´enez, Jos´e Mar´ıa MO5H, 300 MO5H, 350 Mosquera, Alejandro WE1B, 172 Mota, Caroline WE2D, 68 WE2D, 297 Mouslim, Hocine MO2E, 129 Mousseau, Vincent FR2C, 222, 321 FR3E, 246 TH3B, 219 WE2E, 264 Muetterlein, Joschka MO3G, 150 Mula, Josefa TH2D, 143 WE1F, 193 Mustajoki, Jyri MO3E, 279 Myers, Chris MO4G, 176 Nevima, Jan WE2C, 354 Ng, Amos WE1D, 330 MO2F, 99, 168 Nguyen, Viet Hung MO4D, 88 Nielsen, Lars Relund TH2B, 351 Nikulin, Yury MO2B, 348 Nisel, Rauf TU3C, 229 Nisel, Seyhan TU3C, 229 Norese, Maria Franca TH3G, 77 Novello, Chiara TH3G, 77 Novo, Vicente MO4C, 137 Now´e, Ann MO2F, 255 Nowak, Maciej MO2C, 331 Nowak, Uwe TH2E, 307 Ober-Bl¨obaum, Sina TU2G, 360 Ocampo, Germ´an FR3B, 262 Ogryczak, Wlodzimierz WE1G, 324 Ojalehto, Vesa TH2E, 102 Oliveira, Gabriela FR2D, 234 Oliveira, M´onica Duarte MO3C, 260 TU4E, 370 Onsel Ekici, Sule TH4B, 96 Osuna-G´omez, Rafaela MO4C, 171, 374 Ouerdane, Wassila FR2C, 321 WE2E, 264 Ozaydin, Ozay Nasrabadi, Nasim WE1C, 305 Ndiaye, Ismaila Abderhamane TU5G, 261 Nejc, Trdin TU2B, 317 Nekhay, Olexandr TH4B, 120 Neron, Emmanuel TU5G, 261 397 TH4B, 96 Ozcetin, Erdener TH3F, 95 Ozpeynirci, Ozgur MO2C, 115 Ozturk, Gurkan TH2B, 97 TH3C, 187 TH3F, 95 Peng, Hsin-Chuan TU4B, 65 Peng, Kua-Hsin TU4B, 65 Peng, Yi FR2B, 45 Perederieieva, Olga TH4G, 133 Perny, Patrice WE2F, 135 Perzina, Radomir MO3B, 230 Petelin, Dejan TU3F, 203 Pinheiro, Mariana TH4E, 104 Piper, Brian E.B. FR2F, 109 Pirlot, Marc FR2C, 222 MO4F, 173 Pla Santamaria, David TU4G, 299 Podkopaev, Dmitry TH2F, 206 TH2E, 102 Ponte, An´ıbal TH2G, 190 Popovic, Zoran WE2B, 228 Portait, Roland TH3G, 337 Pretolani, Daniele TH2B, 351 Prevost, Aur´elie TU4C, 71 Przybylski, Anthony FR2E, 365 MO4D, 344 Puerto, Justo WE1G, 89 Pulido, Francisco Javier TU5G, 328 P´erez, Carlos Javier MO5H, 47 P´erez, F´atima FR3C, 60 P´erez-de-la-Cruz, Jos´e L. TU5G, 328 P´erez-Gladish, Blanca MO2G, 349 TU4G, 55 TU4G, 299 P´erez-Moreno, Salvador TH4C, 276 P´erez-Rodr´ıguez, Fernando WE1B, 172 P´erez-S´anchez, Danyl TU5C, 67 P´erez-Salas, Jos´e Lu´ıs MO3D, 311 Paoloni, Daniele WE1E, 225 Papamichail, K.Nadia TH3C, 217 Paquete, Luis TH2G, 190 TH2G, 362 Parchment, Avonie TU2C, 107 Patrick, Jonathan TH2C, 181 Pearce, Jon MO2D, 75 Pearman, Alan MO3C, 205 Peidro, David TH2D, 143 WE1F, 193 Pelegr´ın, Blas WE2G, 223 WE2G, 180 Quoc, Ho TH4F, 62 R´ıos Insua, David MO4G, 101 398 R´ıos, John-Jairo FR2D, 368 Raith, Andrea WE2F, 335 TH4G, 48, 133 Ram´ırez, B´arbara MO4F, 240 Ramik, Jaroslav MO3B, 230 Rangaiah, Gade Pandu TH3E, 249 TU5F, 213 Rangel, Luis A.D. FR3E, 123 Ranjithan, S. Ranji FR2F, 109, 326 Ray, Duncan MO3D, 273 Redchuk, Andr´es TU2F, 323 Reichert, Peter TU5C, 208 MO2D, 149 TU4C, 216 Reinerman-Jones, Lauren TU2C, 107 Reunanen, Pasi TH2F, 206 Ribas, Jose Roberto FR3F, 319 MO4E, 221 TH4E, 104 Rocchi, Lucia WE1E, 225 Rocha, Mariana MO4E, 221 Rodr´ıguez D´ıaz, Beatriz TH4C, 276 Rodr´ıguez-Ur´ıa, Mar´ıa Victoria TH2D, 64 Rodriguez, Tinguaro FR3G, 138 Rojo-Alboreca, Alberto WE1B, 172 Rolland, Antoine FR2E, 199 Romero, Carlos TU2B, 157 Rong, Aiying TH4D, 259 Roszkowska, Ewa TH3D, 359 Rubiales , Victoriana MO5H, 136 Rubio, Jos´e Luis FR3G, 162 Rufi´an-Lizana, Antonio MO4C, 374 Rufo, Mar´ıa Jes´ us MO5H, 47 Ruiz, Ana Bel´en MO3F, 291, 355 TU3E, 169 Ruiz, Francisco MO2G, 349 MO3F, 291 TU3E, 169 Ruiz, Francisco Javier TH3C, 245 Ruiz, Juan-Gabriel FR2D, 368 Ruiz-Garz´on, Gabriel MO4C, 374 MO4C, 171 Ruiz-Tagle, Mauricio TH2D, 118 S´anchez, Maria TH3G, 337 S´anchez-Guerrero, Gabriel MO5H, 165 Saaty, Thomas L. FR2B, 142 Sabin, Andrada WE2G, 108 Saborido-Infantes, Rub´en MO5H, 166 MO3F, 355 Salazar-Ordo˜ nez, Melania TU2F, 153 Salda˜ na, Alberto Carlos WE1F, 184 Saleh, Khaled TU4D, 269 Salling, Kim Bang TU5B, 141 Salomon, Valerio 399 WE1B, 338 Salvador, Manuel MO5H, 350 Samavedham, Lakshminarayanan TH3E, 249 Sanchez, Monica TH3C, 245 TU5B, 127 Sankaran, Baskaran MO5H, 254 Sara¸c, Tugba TH4F, 63 TU5E, 144 Sarabando, Paula FR2D, 234 TU2E, 220 Sarin, Rakesh TU4E, 159 Sarkar, Anoop MO5H, 254 Sarkisyan, Rafael MO4G, 329 TU2C, 357 Sarrazin, Renaud TH4C, 158 Savan, Emanuel MO4B, 250 Sayin, Serpil TU5F, 211 TU5F, 147 Sch¨atter, Frank TU5D, 327 Sch¨obel, Anita TH2F, 80 TU5D, 231 WE2F, 335 Schenker, Sebastian WE1G, 98 Schmidt, Marie WE2F, 335 Schnee, Mathias WE2F, 247 Scholten, Lisa FR3F, 237 Schulze, Britta TH2G, 362 Schuwirth, Nele TU4C, 216 TU5C, 208 Segura Garc´ıa del R´ıo, Baldomero MO3D, 311 Segura, Marina MO2D, 304 MO3D, 273 Semenzin, Elena MO5H, 70 WE2C, 72 Sevaux, Marc FR3C, 116 Severo, Juliana Ribas FR3F, 319 MO4E, 221 TH4E, 104 Shalchian, Homayoon MO2G, 73 Shao, Lizhen TH2G, 191 Sharma, Shivom TU5F, 213 Shi, Yong MO2E, 179 Shih, Hsu-Shih TH4B, 94 Shukla, Kalpana TU4D, 270 Siddique, Muhammad Imran FR2D, 186 Siebert, Johannes MO3G, 198 MO3G, 148, 150, 185 Siegmund, Florian MO2F, 99, 168 Silva, Sandra TH4E, 375 Silvennoinen, Kari MO3E, 51 Sindhya, Karthik MO5E, 209 TU3E, 169 TU4F, 81 Siskos, Yannis TU2E, 178 Skocdopolova, Veronika FR3B, 90 Skulimowski, Andrzej M.J. FR2G, 308 400 Slowinski, Roman TH3B, 219 FR2C, 322 TH3B, 272 TU2E, 178 TU3E, 277 TU3F, 369 WE2B, 334 Sobrie, Olivier FR2C, 222 Sohler, Flavio FR3F, 319 Soleimani-Damaneh, Majid TH2C, 373 WE1C, 305 Soletti, Alfredo FR3G, 160 Sousa, Victor FR3F, 318 Soylu, Banu TH2B, 49 WE1C, 103 Spanjaard, Olivier WE2F, 135 Stamati, Ioanna TU2G, 251 Stamenkovic, Mladen WE2B, 228 Stamm, Christian TU4C, 216 Staykov, Boris TH4D, 376 Steponavice, Ingrida TU2G, 253 Steuer, Ralph E. MO2G, 363 Stewart, Theodor TU2F, 267 MO3E, 279 Stidsen, Thomas TU4F, 294 Stiglmayr, Michael TH2G, 190 Stiller, Sebastian TU3E, 316 Sureeyatanapas, Panitas TH4C, 353 Syeda, Darakhshan Jabeen TH3F, 170 Tammer, Christiane TH2F, 80 Tanaka, Tamaki TH2F, 292 Tanino, Tetsuzo TH2F, 292 Taraszewski, Steve TU3B, 93 Tarquis, Ana FR3G, 160, 162, 163 TH3G, 145, 337 Tedeschi, Alessandra MO4G, 101 Teghem, Jacques TH4G, 50 Teich, Jeffrey MO3C, 236 Telen, Dries TU2G, 251 Teo, Grace TU2C, 107 Tervonen, Tommi TH3B, 340 Tezcaner, Diclehan MO4B, 113 Tikkanen, Olli-Pekka TH2F, 206 Timmermann, Robert TU2G, 360 Tlilane, Lydia MO4D, 85 Tomohiro, Hayashida MO3C, 161 TU4E, 258 Topcu, Ilker MO3B, 57 WE1B, 112 Torresan, Silvia MO5H, 70 Toubaline, Sonia MO4G, 243 Toutouh, Jamal WE2E, 295 Trujillo, Tammy FR2D, 368 Trzaskalik, Tadeusz MO2C, 331 401 TU4D, 235 Tsouki`as, Alexis TH3D, 302 Tuˇsar, Tea TU3F, 203 Turhan, Ugur MO4G, 101 Tzeng, Gwo-Hshiung TU4B, 65, 177 MO5H, 367 WE2C, 61 MO4F, 91 Vetschera, Rudolf TU2E, 220 Vidal, Carlos FR3B, 262, 263 Viguri Fuente, Javier MO5H, 53 Villacorta, Pablo TU2D, 194 Villanueva-Rodr´ıguez, Anastasio TH4B, 120 Vincent, Thomas FR2E, 365 TU5C, 293 Vitoriano, Bego˜ na FR3G, 138 TU3C, 78 Udias, Angel TU2F, 323 Ulengin, Fusun TH4B, 96 Urenda Moris, Matias WE1D, 330 Utz, Sebastian MO2G, 363 Wachowicz, Tomasz TH3D, 359 Wallenius, Hannele MO3C, 236 Wallenius, Jyrki MO3E, 51 MO2C, 114 MO3C, 236 Wang, Bo MO2E, 179 Wang, Judith Y. T. MO2D, 75 TH4G, 133 Wen, Hsin-Hui WE2C, 61 Weng, Paul MO4D, 88 Wiecek, Margaret TH2E, 164 Wilson, Reginald TU3D, 232 Wimmer, Maximilian MO2G, 363 Won, Kwang-Jae FR3F, 197 Wong, Wai Keung MO4E, 248 Valle, Rogerio WE1E, 204 Vallerio, Mattia TU2G, 251 Valls, Aida TH4E, 366 WE2B, 334 Van Den Honert, Rob TU5B, 141 van der Vorst, Jack TU3B, 239 Van Impe, Jan TU2G, 251 Van Moffaert, Kristof MO2F, 255 Van Passel, Steven TU5C, 293 Vanderpooten, Daniel FR2E, 285 Vaz, Daniel TH2G, 190 Vercher, Enriqueta WE1F, 196 Verdegay Galdeano, Jos´e Luis TH2D, 332 Vergara-Maldonado, E. Ivonne MO5H, 165 Verlinde, Sara Xianhua, Wei MO2E, 179 Xie, Naiming 402 MO3C, 205 Xu, Ling MO3E, 182 WE2G, 108 MO4B, 250 MO4E, 121 MO5H, 218 WE2C, 110 Yılmaz, Hafize WE1B, 112 Yamada, Syuuji TH2F, 292 Yang, Jian-Bo MO3E, 182 MO4B, 250 MO4E, 121 TH4C, 353 Yepes, V´ıctor TU5E, 268 Yildiz, Gazi Bilal WE1C, 103 Zabeo, Alex MO5H, 70 WE2C, 72 Zahedi, Siamak TU5B, 127 Zaroliagis, Christos TU5G, 174 Zheng, Jun TU3G, 215 Zielniewicz, Piotr WE2B, 334 Zietsman, Joe TH4C, 346 Zilinskas, Antanas TH4D, 286 Zilinskas, Julius TH4D, 286 WE2G, 180 403

© Copyright 2019