A consensus model for group decision-making with personalized individual self-confidence and trust semantics: A perspective on dynamic social network interactions

Abstract

Natural language could be a common method for individuals to specific social relationships or self-evaluation (i.e., assured or unconfident). Hence, it is important to contemplate customized linguistics in collective decisions that involve linguistic variables, because the same linguistic term set may have totally different individual numerical scales. This research develops a consensus model taking into consideration the personalized individual self-confidence and trust semantics in dynamic social network group decision-making scenarios. Firstly, an optimization model driven by best additive consistency is proposed to derive the personalized self-confidence and trust semantics of individuals. Subsequently, a trust aggregation technique supporting the shortest path is outlined for weight distribution. The proposed technique not only considers the personalized individual trust linguistics but can also effectively decrease the risk of information distortion caused by trust decay. Moreover, a consensus feedback strategy is developed that combines opinion adjustment and dynamic trust interaction to enhance the potency of consensus. Totally different from most existing consensus models supported static trust network, the bestowed consensus feedback mechanism considers the inverse driving result of opinion adjustment on trust interaction within the consensus reaching processes. Finally, numerical examples and comparisons are provided to justify the practicability of the proposed model.