Large-scale consensus with dynamic trust and optimal reference in social network under incomplete probabilistic linguistic circumstance

Abstract

Consensus plays a pivotal role in large-scale group decision-making, and the social network relationship significantly impacts the consensus reaching process (CRP). Consequently, there is considerable focus on achieving large-scale consensus within social network (LSC-SN). Typically, experts are used to expressing their evaluations with language that can be effectively characterized by probabilistic linguistic term set (PLTS). However, the existing distance measure of PLTS exhibits drawbacks, and PLTSs provided by experts might be incomplete. To address the two challenges, this research defines an improved distance measure and proposes an estimation method for incomplete PLTS from the perspectives of trust relationships, collaborative filtering, and confidence level. Subsequently, in the CRP, a hierarchical clustering algorithm based on the trust-similarity measure is developed to cluster experts who share mutual trust and similar evaluations. Then, a consensus model employing dynamic trust and optimal reference is introduced to achieve both intra-cluster and inter-cluster consensus simultaneously. Finally, to illustrate the feasibility and advantage of the proposed model, a numerical experiment and the comparative analyses are conducted.