Two-Stage Heterogeneous Noncooperative Behavior for Consensus in Large-Scale Group Decision Under Fuzzy Distributed Preference Relation

Large-scale multiattribute group decision making (LS-MAGDM) is common in practice. How to apply it and form a consistent solution has become a hot topic in decision science research. The evaluation information expression and experts’ behaviors formed during decision-making process have been shown to have important impact for reaching consensus. Based on this, consensus reaching process for LS-MAGDM considering opinion leaders and two-stage heterogeneous noncooperative behavior is proposed. First, we define cardinal consistency of fuzzy distributed preference relation (FDPR) based on additive consistency. The FDPR score calculation method satisfying additive consistency is presented. Second, we put forward the identification method of opinion leaders. Starting from social network and evaluation information, the concepts of authority, stability, professionalism, and reliability are defined and combined to measure expert’s leadership ability. Then, the multiobjective optimization model is constructed to identify opinion leader in each subgroup. Finally, consensus reaching models based on two-stage heterogeneous noncooperative behavior are proposed. noncooperative degree is introduced to measure noncooperative behavior of experts or subgroups. Different feedback mechanisms and noncooperative behavior management methods are applied to both the intra-subgroup and inter-subgroup consensus reaching stages. Finally, the feasibility of this method is verified by a case study of remanufacturer quality assessment and comparative analysis with some State-of-the-Art methods.