Modeling interactions among criteria in MCDM methods: A review

Abstract

This paper reviews approaches for modelling interactions and dependencies between criteria in multi-criteria decision-making (MCDM) methods. Traditionally, MCDM methods only allow the establishment of linear dependence between criteria, so they only allow for simplified models that are mostly inadequate for modelling real-life problems. Several methods have therefore been developed for modelling the interdependence between criteria and sub-criteria in MCDM. This paper makes a comparison between some popular methods that allow modelling of criteria interdependencies: Analytic Network Process (ANP), Decision Making Trial and Evaluation Laboratory (DEMATEL), Interpretive Structural Modelling (ISM), Fuzzy Measures and the Choquet Integral (CI) and Interpolative Boolean Algebra (IBA). These methods allow the establishment of interactions and comparisons between criteria using supermatrices, diagrams, fuzzy measures, fuzzy integrals and logical functions. This paper presents the MCDM approaches that include in their analysis the interdependencies and relation between criteria/sub-criteria and thus enable more efficient and realistic modelling of decision-making problems.