Pseudo-trust behavior-based consensus methodology with asymmetric costs for social network group decision-making

Abstract

In intelligent decision-making systems, the coupling effect of information fuzziness with multiple constraints-including social network restrictions, pseudo-trust behavior interference, and fairness imbalance, not only amplifies the positional divergences among decision-makers but also significantly exacerbates the practical dilemma of alternative selection. To address these issues, this paper develops a social network-based group consensus optimization method. It effectively overcomes the multiple constraints in the consensus reaching process by integrating a pseudo-trust identification and calibration mechanism, while balancing the interest demands among decision-makers, thereby providing a feasible solution for group decision-making. Firstly, a social network topology generation method integrating fairness criteria, opinion similarity measurement, opinion credibility evaluation, and conflict quantification is developed by deconstructing the coupling mechanism of multi-dimensional characteristic factors in social networks. On this basis, the opinion evolution trajectory of decision-makers during the consensus negotiation process is accurately characterized, and an optimization model accounting for the asymmetric effect of bidirectional opinion adjustment is further embedded to achieve efficient convergence of group consensus. Upon the completion of consensus reaching, the deviation magnitude between decision-makers’ consensus opinions and their initial opinions, as well as the dynamic evolution differences in consensus levels, are comparatively analyzed to identify the opinion evolution path deviation induced by pseudo-trust behaviors. Based on the theoretical framework of three-way decisions, refined classified control is then implemented for different types of opinion evolution deviations. Finally, a set of real-world medical decision-making data is used to conduct case analysis and comparative experiments, which verifies the superiority and stability of the developed method.