Enriched preference modeling and robustness analysis for the ELECTRE Tri-B method

We consider multiple criteria sorting problems with preference-ordered classes delimited by a set of boundary profiles. While significantly extending the ELECTRE Tri-B method, we present an integrated framework for modeling indirect preference information and conducting robustness analysis. We allow...

Full description

Saved in:
Bibliographic Details
Published in:Annals of operations research Vol. 306; no. 1-2; pp. 173 - 207
Main Authors: Kadziński, Miłosz, Martyn, Magdalena
Format: Journal Article
Language:English
Published: New York Springer US 01.11.2021
Springer
Springer Nature B.V
Subjects:
Analysis
Business and Management
Combinatorics
Consumer preferences
Decision analysis
Decision making
Mathematical programming
Multiple criteria decision making
Multiple criterion
Operations research
Operations Research/Decision Theory
Preferences
Robustness (mathematics)
S.I.: Regression Methods based on OR techniques
Theory of Computation
Poland
Preference modeling
Multiple criteria sorting
Multiple criteria decision aiding
Outranking relation
Robustness analysis
Monte Carlo simulation
ISSN:0254-5330, 1572-9338
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We consider multiple criteria sorting problems with preference-ordered classes delimited by a set of boundary profiles. While significantly extending the ELECTRE Tri-B method, we present an integrated framework for modeling indirect preference information and conducting robustness analysis. We allow the Decision Maker (DM) to provide the following three types of holistic judgments: assignment examples, assignment-based pairwise comparisons, and desired class cardinalities. A diversity of recommendation that can be obtained given the plurality of outranking-based sorting models compatible with the DM’s preferences is quantified by means of six types of results. These include possible assignments, class acceptability indices, necessary assignment-based preference relation, assignment-based outranking indices, extreme class cardinalities, and class cardinality indices. We discuss the impact of preference information on the derived outcomes, the interrelations between the exact results computed with mathematical programming and stochastic indices estimated with the Monte Carlo simulations, and new measures for quantifying the robustness of results. The practical usefulness of the approach is illustrated on data from the Financial Times concerning MBA programs.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0254-5330
1572-9338
DOI:10.1007/s10479-020-03833-z