Dynamic consensus of large group emergency decision-making under dual-trust relationship-based social network

Abstract

Trust relationship is an effective basis for reducing the dimensionality of decision makers (DMs) for group decision making (GDM) problems in the context of social networks, and have received extensive attention in the past few years. However, most existing studies describe the trust relationship only from the single perspective of the interpersonal relationships between DMs, which easily leads to inaccurate decision-making results. In this paper, we propose a dynamic consensus method using a dual-trust relationship-based social network (DTRSN) to solve large group emergency decision-making (LGEDM) problems. First, to more accurately reflect the trust relationship among DMs, a DTRSN based on familiarity-based trust and similarity-based trust is developed, which can comprehensively model the interpersonal relationships and opinion similarity between DMs simultaneously. Next, a DTRSN-based clustering method that considers dual-trust relationships is introduced to divide the large-group DMs into several clusters. Subsequently, to improve the efficiency of LGEDM, the exit-delegation mechanism is modified and introduced into our study, which can speed up consensus by excluding some subgroups whose opinions differ greatly from other subgroups and reserve their influence. Subsequently, a dynamic consensus method based on the modified exit-delegation mechanism is proposed to help DMs reach an agreement quickly. Finally, a case study regarding emergency alternative selection to deal with heavy rainstorms in Zhengzhou, China, is conducted to verify the efficiency of our study, and several comparison and simulation analyses are also conducted to demonstrate the feasibility and practicality of the proposed method.