Constrained Community Detection and Multistage Multicost Consensus in Social Network Large-Scale Decision-Making

As a widely encountered decision-making scenario in modern society, social network large-scale decision-making (SNLSDM) is becoming a frontier topic in the field of decision science. Decision information usually includes social relationships among decision-makers (DMs) and individual opinions. This makes clustering and consensus building, which are two important processes for solving SNLSDM problems, significantly complex and requires the integration of multiple factors. This study designs a decision support method that consists of a constrained community detection (CCD) method and a multistage multicost consensus (MSMulCC) model. The CCD method takes the similarities among individual opinions as the mandatory constraint to guide the classification of DMs based on social relationships. The consensus-reaching process (CRP) is an effective tool for reducing differences of opinion. We hold that the DM with high compatibility but low consensus can reduce the adjustment amount by actively losing some compatibility. In this way, three types of consensus costs are generated, including individual adjustment cost, group adjustment cost, and compatibility loss cost. In this study, an MSMulCC model is developed, and the impact of different types of consensus costs on CRP. Finally, the feasibility and characteristics of the proposal are revealed through a case study and comparative analysis of the supply chain financing model selection.