A novel multi-attribute group decision-making method for heterogeneous preferences and attribute interaction contexts. (1st July 2023)

• Record Type: Journal Article
• Authors: Zhao, Chengwei; Xu, Xuanhua; Liu, Ruihuan; He, Jishan
• Is Part Of: Expert systems with applications. Volume 221(2023)
• Publication Date: 2023-07-01
• Subjects: Multi-attribute group decision-making -- Heterogeneous preferences -- Attribute interaction -- Shapley-based dynamical variable weights -- Choquet integral; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2023.119752
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A TD-DU space method is presented to homogenize heterogeneous preferences. The SDVW is defined to determine the interaction attribute weight. An new SDVW-Choquet integral is established to fuse non-additive preferences. A MAGDM method for heterogeneous preferences and attribute interaction is proposed. Abstract: Heterogeneous preferences and attribute interaction are two important features in realistic multi-attribute group decision-making (MAGDM), directly affecting the decision quality. However, few MAGDM methods can handle both. This study proposes a novel MAGDM method suitable for heterogeneous preferences and attribute interaction environments. First, a three-dimensional determinacy-uncertainty space method for homogenizing the heterogeneous preferences of real numbers, interval numbers, triangular fuzzy numbers, and triangular intuitionistic fuzzy numbers is proposed, avoiding information loss while having a simple homogeneous data form. Considering that the static weight method commonly used in existing MAGDM cannot fully reflect the variation of attribute importance for different alternatives, the Shapley-based dynamic variable weight (SDVW) model is innovatively constructed to determine the interaction attribute weights. Subsequently, the SDVW-Choquet integral is established to fuse the non-additive preferences, solving the compensation problem of non-additive preferences fusion and simplifying the complexity of the fusion process. Finally, a case study and comparison analyses demonstrate the effectiveness, robustness, and advantages of the proposed method. The results show that simultaneously considering heterogeneous preferences and attribute interaction can improve the accuracy and reliability of the MAGDM ranking results.