Social network-based overlapping community clustering and feedback mechanism for large-scale group decision making

In large-scale group decision-making (LSGDM) problems, decision-makers (DMs) often exhibit multi-benefit properties, leading to overlapping communities where a DM belongs to multiple subgroups. However, most existing studies assume that a DM belongs to only one subgroup, with little consideration given to overlapping communities in LSGDM events. To address this gap, this paper proposes an overlapping community clustering algorithm in LSGDM based on trust relationships generated by collaborative representation (CR-OCA) to cluster DMs and detect overlapping DMs. Studies that consider the phenomenon of overlapping communities primarily emphasize the positive role of overlapping DMs in facilitating consensus, but they may also have negative influences on the consensus reaching process (CRP) in real situations. To address this complexity, the overlapping community-based consensus reaching process (OC-CRP) is proposed, which comprehensively considers the trust-conflict social network among DMs and positive or negative effects of overlapping DMs on consensus. Given the diverse structures of overlapping communities and their varying impacts on CRP, a feedback mechanism with tailored adjustment rules for different overlapping structures is developed. Finally, numerical examples and comparative analyses are used to show the feasibility and effectiveness of the proposed CR-OCA and OC-CRP, demonstrating their superiority in enhancing clustering stability and managing complex overlapping communities in LSGDM.