Intelligent Decision Support Systems: Combining Operations Research and Artificial Intelligence - Essays in Honor of Roman Słowiński (Multiple Criteria Decision Making)

Preface
Acknowledgments
Contents
Contributors
1 Roman Słowiński and His Research Program: Intelligent Decision Support Systems Between Operations Research and Artificial Intelligence
	1.1 Introduction
	1.2 Short Biographical Notes
	1.3 Advice and Supervision of New Researchers
	1.4 The Main Research Contributions of Roman Słowiński
	1.5 A Bibliometric Analysis of the Research Activity and Impact of Roman Słowiński
		1.5.1 Numbers of Publications and Citations
		1.5.2 The Most Cited Papers
		1.5.3 Journals and Other Sources of Publications
		1.5.4 Main Co-Authors
		1.5.5 Most Contributed Topics
	1.6 Volume's Contributions
	1.7 Final Remarks
	References
2 Roman's Scientific Trajectory: A Retrospective with an Emphasis on the Beginning
	References
3 ELECTRE Methods: A Survey on Roman SłowińskiContributions
	3.1 Introduction
	3.2 A Brief Overview on ELECTRE Main Concepts
	3.3 Roman Słowinski Contributions
		3.3.1 ``Inferring an ELECTRE Tri Model from Assignment Examples'' and ``A User-Oriented Implementation of the ELECTRE-TRI Method Integrating Preference Elicitation Support''
		3.3.2 Searching for an Equivalence Between Decision Rules and Concordance-Discordance Preference Model in Multicriteria Choice Problems
		3.3.3 Axiomatization of Utility, Outranking, and Decision Rule Preference Models for Multiple Criteria Classification Problems Under Partial Inconsistency with the Dominance Principle
		3.3.4 Handling Effects of Reinforced Preference and Counter-Veto in Credibility of Outranking
		3.3.5 ELECTREGKMS: Robust Ordinal Regression for Outranking Methods
		3.3.6 Multiple Criteria Hierarchy Process with ELECTRE and PROMETHEE
		3.3.7 Multiple Criteria Hierarchy Process for ELECTRE Tri Methods
		3.3.8 A Robust Ranking Method Extending ELECTRE III to Hierarchy of Interacting Criteria, Imprecise Weights and Stochastic Analysis
		3.3.9 ELECTRE-III-H: An Outranking-Based Decision Aiding Method for Hierarchically Structured Criteria
		3.3.10 Other Contributions
	3.4 Conclusions
	References
4 How Can Decision Sciences and MCDM Help Solve Challenging World Problems?
	4.1 Introduction
	4.2 Internet Searches
		4.2.1 Recommender Systems
	4.3 Big Data (and Artificial Intelligence)
	4.4 The Platform Economy
	4.5 Climate Change, Concern for Environment
	4.6 Where the Opportunities Lie?
	4.7 Final Word
	References
5 Preference Disaggregation Analysis: An Overview of Methodological Advances and Applications
	5.1 Introduction
	5.2 The General Framework of Preference Disaggregation Analysis
	5.3 Models, Formulations, and Methodological Advances
		5.3.1 Value Function Models
			5.3.1.1 Modeling Forms
			5.3.1.2 Ranking Problems
			5.3.1.3 Sorting/Classification Problems
			5.3.1.4 MUSA Method
		5.3.2 Outranking Models
			5.3.2.1 Brief Outline of Outranking Relations
			5.3.2.2 Inference Procedures
		5.3.3 Rule-Based Models
	5.4 The Robustness Concern
		5.4.1 The Issue of Robustness in PDA
		5.4.2 Methodological Approaches
		5.4.3 Experimental Studies
	5.5 Overview of Applications and Decision Support Systems
	5.6 Conclusions and Future Perspectives
	References
6 Modeling and Learning of Hierarchical Decision Models: The Case of the Choquet Integral
	6.1 Introduction
	6.2 Hierarchical Multi-Criteria Decision Models
		6.2.1 Multi-Criteria Decision Models
		6.2.2 Choquet Integral
		6.2.3 Hierarchical Multi-Criteria Decision Models
	6.3 Expressiveness of Hierarchical Models: The Case of Choquet Integrals
		6.3.1 An Illustrative Example in an MCDA Setting
		6.3.2 The Hierarchical Choquet Integral Model
	6.4 Neur-HCI Framework to Learn a HMCDM
		6.4.1 Marginal Utility Functions
		6.4.2 Learning Tasks
		6.4.3 Learning and Optimization
	6.5 Summary and Conclusion
	References
7 Preference Learning Applied to Credit Rating: Applications and Perspectives
	7.1 Introduction
	7.2 Credit Rating or Sorting with Multiple Criteria
		7.2.1 Country Risk
		7.2.2 Corporate Credit Risk
	7.3 Applications
		7.3.1 Sorting Sovereign Bonds with Two Preference Learning Approaches
			7.3.1.1 Alternatives and Reference Set Construction
			7.3.1.2 Conditional Criteria and Data Reduction
			7.3.1.3 Preference Learning Approach
		7.3.2 Sorting Brazilian Debentures with the PDTOPSIS-Sort Method
			7.3.2.1 Problem Definition
			7.3.2.2 Learning the Expert's Preferences
	7.4 Conclusion and Future Perspectives
	References
8 USort-nB and USort-nC: Two Multi-criteria Ordinal Classification Methods Using Interval Value Functions
	8.1 Introduction
	8.2 Some Background
	8.3 An Ordinal Classification Method Based on Limiting Boundary Actions
		8.3.1 Description of the Method
		8.3.2 Consistency Properties of the USort-nB Primal and Dual Procedures
	8.4 USort-nC: An Ordinal Classification Method Based on Representative Actions
		8.4.1 Description of the Method
		8.4.2 Consistency Properties of USort-nC
		8.4.3 An Illustrative Example
	8.5 Concluding Remarks
	References
9 Constructing an Outranking Relation from Semantic Criteria and Ordinal Criteria for the ELECTRE Method
	9.1 Introduction
	9.2 An Introduction to the Classic Methodology for Constructing an Outranking Relation in ELECTRE
	9.3 Constructing an Outranking Relation from Semantic Criteria
	9.4 Constructing an Outranking Relation from a Partial Pre-order on a Set of Alternatives
	9.5 Constructing an Outranking Relation from an Assignment of Alternatives to Categories
	9.6 The ELECTRE-H Software Package: Tools for Assisting the Decision Maker in the Determination of Thresholds
	9.7 Examples of Applications
		9.7.1 Environmental Analysis
		9.7.2 Tourism
	9.8 Summary and Conclusions
	References
10 Robust Ordinal Regression for Multiple Criteria Decision Aiding
	10.1 Introduction
	10.2 Review of Core Methods in Robust Ordinal Regression
		10.2.1 Problem Typologies Considered in the ROR Methods
		10.2.2 Preference Information Elicited in the ROR Methods
		10.2.3 Preference Models Employed in the ROR Methods
		10.2.4 Decision Outcomes Provided in the ROR Methods
	10.3 Review of Other Developments Related to Robust Ordinal Regression
		10.3.1 Robust Ordinal Regression methods
		10.3.2 Decision Aiding Methods Related to ROR
		10.3.3 Real-World Applications
	10.4 Summary
	References
11 What Is Legitimate Decision Support?
	11.1 Introduction
	11.2 The Legitimacy of Decision Support: An Important But Neglected Topic
	11.3 Visions of Legitimacy
	11.4 The Legitimacy of Decision Support: A General Theory
	11.5 Hurdles on the Road to Legitimacy
	11.6 Conclusions
	References
12 MR-Sort with Partial Information to Decide Whether to Invest in Innovation Projects
	12.1 Introduction
	12.2 Multiple Criteria Sorting When Evaluations Are Missing
	12.3 MR-Sort with Partial Information
		12.3.1 A Short Reminder of MR-Sort
		12.3.2 Our Proposal
		12.3.3 The Bipolar Ordered Set of Classes
		12.3.4 The Coalition Weights
		12.3.5 The Conservative Mindset for Sorting
		12.3.6 The Audacious Mindset for Sorting
	12.4 Application to the Evaluation of Innovation Projects
	12.5 Conclusion and Discussion
	References
13 Meta-Rankings of Journals Publishing Multiple Criteria Decision Aiding Research: Benefit-of-Doubt Composite Indicators for Heterogeneous Qualitative Scales
	13.1 Introduction
	13.2 Benefit-of-Doubt Perspectives for Heterogeneous Qualitative Scales
		13.2.1 Notation and Common Elements
		13.2.2 Model 1: BoD Minimize Regret
		13.2.3 Model 2: BoD Minimize Regret without Trade-Offs
		13.2.4 Model 3: BoD Maximum Value
	13.3 Selection of MCDA-Related Journals and Rankings
	13.4 Results for the Different Models and Variants
		13.4.1 Model Comparison
		13.4.2 Variant Comparison
		13.4.3 Journal Comparison
	13.5 Summary
	References
14 Interactive Multicriteria Methodology Based on a Synergy of PROMETHEE II and Robust Simos Methods: Application to the Evaluation of E-government in Europe
	14.1 Introduction
	14.2 The Synergy of PROMETHEE II and Robust Simos Multicriteria Evaluation Methods
		14.2.1 A Brief Presentation of the PROMETHEE Methods
			14.2.1.1 A Multicriteria Pairwise Outranking Indicator
			14.2.1.2 Outranking Flows
			14.2.1.3 PROMETHEE II Ranking Procedure
		14.2.2 Elicitation of Criteria Weights Through the Simos Methods
			14.2.2.1 The Original Method of Cards
			14.2.2.2 The Robust Simos Method
	14.3 Bipolar Robustness Control and Decision Support
		14.3.1 Methodological Framework of Bipolar Robustness Control
		14.3.2 Robustness Measures in the Disaggregation Pole
		14.3.3 Robustness Measures in the Aggregation Pole and Decision Support
	14.4 Implementation to the E-government Evaluation: Phase A'—First Elicitation of the Criteria Weights
		14.4.1 E-government Evaluation Importance and Criteria Modelling
		14.4.2 Initialization of the PROMETHEE II Method
		14.4.3 A΄ Phase of the Robust Simos Method
	14.5 Implementation to the E-government Evaluation: B and C Phases
		14.5.1 B Phase of Robustness Control
		14.5.2 C Phase of Robustness Control
	14.6 Conclusion
	Appendices
		Appendix 1: Typology of PROMETHE's Generalized Criteria: Preference Function P(d),d: Evaluation Difference [2]
		Appendix 2: Performance of the European Countries on the Eight Evaluation Criteria
	References
15 The Use of Decision Maker's Preferences in Multiobjective Metaheuristics
	15.1 Introduction
	15.2 Two-Stage Approach for Interactive Analysis of Multiobjective Combinatorial Optimization Problems
	15.3 Preference-Based Evaluation of Multiobjective Metaheuristics
	15.4 Evolutionary Multiobjective Optimization Algorithms Based on Robust Ordinal Regression
	15.5 Conclusion
	References
16 Decomposition and Coordination for Many-ObjectiveOptimization
	16.1 Introduction
	16.2 Foundations for Decomposition
	16.3 State-of-the-Art in Multiobjective Decomposition and Coordination
		16.3.1 Models and Solution Concepts
		16.3.2 Coordination and Decision-Making
		16.3.3 Applications
	16.4 Developing a Decomposition–Coordination Technique
		16.4.1 Types of Decomposition
		16.4.2 Efficient Solutions of Subproblems Up to AiO
		16.4.3 Efficient Solutions of AiO Down to Subproblems
		16.4.4 Coordinating Subproblems
		16.4.5 Computation of Efficient Solutions
		16.4.6 Interactive Process
	16.5 Example
	16.6 Conclusions
	References
17 Fuzzy Linear Programming with General Necessity Measures
	17.1 Introduction
	17.2 Necessity Measures
	17.3 Possibilistic Linear Programming
	17.4 Case Where θi(·,hi)'s Are Convex and Concave
	17.5 Similar Results for Necessity Measures Defined by Modifier-Generating Functions
	17.6 Results Applied to Various Implication Functions
		17.6.1 R-, Reciprocal R-, and S-Implication Functions
		17.6.2 Results in R-Implication Functions
		17.6.3 Results in Reciprocal R-Implication Functions
		17.6.4 Results in S-Implication Functions
		17.6.5 Obtained Results Applied to Famous Implication Functions
	17.7 Concluding Remarks
	References
18 Dominance-Based Rough Set Approach: Basic Ideas and Main Trends
	18.1 Introduction
	18.2 Some Historical Notes on the Dominance-Based Rough Set Approach
	18.3 Basic Concepts of the Dominance-Based Rough Set Approach
	18.4 Developments of DRSA
		18.4.1 DRSA to Multicriteria Choice and Ranking
		18.4.2 Case-Based Reasoning Using Dominance-Based Decision Rules
		18.4.3 Adaptations of DRSA to Handle Missing Attribute Values
		18.4.4 Extensions of DRSA for Interval Evaluations on Criteria
		18.4.5 Extending DRSA to Address Hierarchical Structure of Attributes
		18.4.6 Extensions of DRSA for Non-ordinal Data
		18.4.7 Parametric, Decision Theoretic, and Stochastic DRSA
		18.4.8 Decision Rules Induction
	18.5 Available Software
		18.5.1 ruleLearn
		18.5.2 RuLeStudio
		18.5.3 RuleVisualization
		18.5.4 jMAF
		18.5.5 RuleRank Ultimate Desktop Edition
	18.6 Conclusions
	Acknowledgements
	References
19 Rule Set Complexity for Mining Incomplete Data Using Probabilistic Approximations Based on Generalized Maximal Consistent Blocks
	19.1 Introduction
	19.2 Incomplete Data
	19.3 Probabilistic Approximations
		19.3.1 Global Probabilistic Approximations
		19.3.2 Saturated Probabilistic Approximations
		19.3.3 Rule Induction
	19.4 Experiments
	19.5 Conclusions
	References
20 Rule Confirmation Measures: Properties, Visual Analysis and Applications
	20.1 Introduction
	20.2 Bayesian Confirmation and Rule Confirmation Measures
		20.2.1 Bayesian Confirmation
		20.2.2 Rule Confirmation Measures
	20.3 Properties of Confirmation Measures
		20.3.1 Property of Monotonicity M
		20.3.2 Symmetry Properties
		20.3.3 Properties Inspired by Extreme Values of Confirmation
	20.4 Visual Analysis of Measure Properties
		20.4.1 Characteristic Regions of Confirmation Measures
		20.4.2 Monotonicity-Related Properties: M
		20.4.3 Symmetry-Related Properties: ES, HS, and EHS
		20.4.4 Extrema-Related Properties: L, Ex1, and Maximality/Minimality
	20.5 Current and Future Applications to Data Mining and Machine Learning
		20.5.1 Multi-Criteria Selection of Non-dominated Rules
		20.5.2 Post-processing of Rules Induced from Imbalanced Data
		20.5.3 Classification by Association and Biomedical Data
		20.5.4 Tetrahedron Visualizations for Classifier Evaluation Measures
		20.5.5 Future Directions
	References
21 An Approach to Combining Adherence-to-Therapy and Patient Preference Models for Evaluation of Therapies in Patient-Centered Care
	21.1 Introduction
	21.2 Background
	21.3 Atrial Fibrillation
	21.4 Materials and Methods
		21.4.1 Adherence Profile—Identifying Criteria
		21.4.2 Adherence Profile—Selecting an Adherence Category
		21.4.3 Adherence Model—Identifying Criteria
		21.4.4 Adherence Model—Capturing Preferential Information Regarding Therapies
		21.4.5 Patient Preference Models
	21.5 Calculations
	21.6 Results and Discussion
	21.7 Conclusions
	References