Consistency and consensus reaching process for group decision making based on complete interval distributed preference relations under social network analysis

• Published in: Information fusion Vol. 88; pp. 126 - 145
• Main Authors: Zhou, Mi, Guan, Zhong-Xu, Chen, Yu-Wang, Zhou, Zhi-Ping, Wu, Jian, Herrera-Viedma, Enrique
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.12.2022
• Subjects: Complete interval distributed preference relation; Consensus reaching process; Consistency; Distributed linguistic trust relation; Group decision making; Opinion fusion; Group decision making; Distributed linguistic trust relation; Consensus reaching process; Complete interval distributed preference relation; Opinion fusion; Consistency
• ISSN: 1566-2535; 1872-6305
• DOI: 10.1016/j.inffus.2022.07.015

•The consistency of complete interval distributed preference relation (CIDPR) is defined.•Adjustment algorithm is proposed to convert inconsistent CIDPR to be acceptable consistent level.•Experts’ weights are generated based on DLTRs by the ER algorithm.•Consensus measurement and adjustment methods under social network analysis are proposed. Group decision-making (GDM) problems often consist of many indeterminacy factors in realistic situation. How to cope with consistency and consensus under uncertain circumstance are two critical issues in pairwise comparison based GDM problems. In this paper, we firstly propose the model of complete interval distributed preference relation (CIDPR) based on the concept of linguistic distribution with interval symbolic proportions, distribution linguistic preference relation (DLPR) and IDPR. Secondly, the additive consistency index of CIDPR is defined to measure the consistency level of expert's judgment, and an adjustment algorithm is proposed for converting inconsistent CIDPR to an acceptable consistent level. Thirdly, since trust relation is a critical factor in the generation of experts’ weights and the adjustment of experts’ opinions, consensus reaching process (CRP) is designed to take into account distributed linguistic trust relations under social network analysis (SNA). In the proposed adjustment mechanism, non-consensus individual should modify opinion towards his/her trusted and highly weighted expert. The advantage of the proposed inconsistent CIDPR adjustment model can maximally retain the information in the original distribution, while the CRP has a relatively fast convergent speed and good practicality. An illustrative example of strategic new product selection is conducted to demonstrate the applicability of the proposed method and its potential in supporting realistic GDM problems.