Many-Criteria Optimization and Decision Analysis State-of-the-Art, Present Challenges, and Future Perspectives

This book presents the state-of-the-art, current challenges, and future perspectives for the field of many-criteria optimization and decision analysis. The field recognizes that real-life problems often involve trying to balance a multiplicity of considerations simultaneously – such as performance, cost, risk, sustainability, and quality. The field develops theory, methods and tools that can support decision makers in finding appropriate solutions when faced with many (typically more than three) such criteria at the same time. 

The book consists of two parts: key research topics, and emerging topics. Part I begins with a general introduction to many-criteria optimization, perspectives from research leaders in real-world problems, and a contemporary survey of the attributes of problems of this kind. This part continues with chapters on fundamental aspects of many-criteria optimization, namely on order relations, quality measures, benchmarking, visualization, and theoretical considerations. Part II offers more specialized chapters on correlated objectives, heterogeneous objectives, Bayesian optimization, and game theory.

Written by leading experts across the field of many-criteria optimization, this book will be an essential resource for researchers in the fields of evolutionary computing, operations research, multiobjective optimization, and decision science.

Chapter 1: Introduction to Many-Criteria Optimization and Decision Analysis.- Chapter 2: Key Issues in Real-World Applications of Many-Objective Optimisation and Decision Analysis.- Chapter 3: Identifying Properties of Real-World Optimisation Problems through a Questionnaire.- Chapter 4: Many-Criteria Dominance Relations.- Chapter 5: Many-Objective Quality Measures.- Chapter 6: Benchmarking.- Chapter 7: Visualisation for decision support in many-objective optimisation: state-of-the-art, guidance and future directions.- Chapter 8: Theoretical Aspects of Subset Selection in Multi-Objective Optimisation.- Chapter 9: Identifying Correlations in Understanding and Solving Multi-Objective Problems.- Chapter 10: Bayesian Optimization.- Chapter 11: A game theoretic perspective on Bayesian many-objective optimization.- Chapter 12: Heterogeneous Objectives: State-of-the-Art and Future Research.- Chapter 13: MACODA Ontology and Knowledge Management.

Dimo Brockhoff is a researcher at Inria, France, and member of the Randomized Optimization team that is co-located with Ecole Polytechnique, IP Paris. He studied computer science in Dortmund, Germany and received a PhD from ETH Zurich, Switzerland for his work “Many-Objective Optimization and Hypervolume Based Search”. His research interests are focused on multiobjective optimization and benchmarking.

Michael Emmerich is an Associate Professor at Leiden University and has recently also been an international research fellow at the University of Jyväskylä, Finland. At Leiden University he leads since 2011 the Multi-objective Optimization and Decision Analysis (MODA) research group. He received his doctorate in 2005 on the topic "Gaussian Process Models in Multiobjective Optimization" from TU Dortmund, Germany (promotor: Hans-Paul Schwefel). He chaired four leading international optimization conferences and five Lorentz center workshops. He published more than 200articles on multiobjective optimization, mainly indicator-based algorithms and Bayesian multiobjective optimization, and applications such as architectural design, logistics, and computational chemistry.

Robin Purshouse is Professor of Decision Sciences at the University of Sheffield, UK. He received his PhD in Control Systems from the University of Sheffield in 2004 for his thesis “On the Evolutionary Optimisation of Many Objectives”, which was one of the earliest works on the topic of many-criteria optimization. He was General Chair of the Seventh International Conference on Evolutionary Multi-Criterion Optimization held in Sheffield in 2013. His research interests in multi-objective optimization are inspired by real-world problems and include robust optimization, surrogate-based optimization, and multi-disciplinary optimization.