A similarity-conformity-based clustering and decentralized feedback mechanism to manage incompatible decision-makers in large-scale group decision-making

Abstract

This study proposes a similarity-conformity-based clustering and decentralized feedback mechanism to manage incompatible decision-makers (DMs) in the large-scale group construction decision-making problem. Firstly, conformity is proposed to measure the consistency of the most preferred alternative of any two DMs and then extends this concept to define the similarity-conformity level of DMs. The average conformity between DMs in a subgroup represents the cohesion degree of this subgroup. Based on these concepts, a new clustering method considers the conflicts of interest among DMs and introduces the similarity-conformity to divide them into several subgroups. Then, the number of DMs in a subgroup and the cohesion degree are combined to formulate the subgroups’ weights. A decentralized feedback mechanism is proposed to manage two categories of incompatible DMs in the consensus-reaching process. The proposed consensus model adjusts the incompatible DMs’ opinions or subgroups’ weights under four scenarios. In the proposed model, the four adjustment methods prove to be convergent, and they improve the group consensus level by ensuring the DMs’ conformity and the logical consistency of preference relation. Finally, it provides an example of the alternative selection for transportation infrastructure construction. The comparative study and the sensitivity analysis are performed to verify the effectiveness of the proposed method.