An opinion evolution-based consensus-reaching model for large-scale group decision-making: Incorporating implicit trust and individual influence

Abstract

Large-scale group decision-making (LGDM) often involves complex challenges, such as effectively clustering decision-makers, modeling asymmetric trust relationships, and balancing the influence of leaders and members to achieve consensus. This study proposes a novel opinion evolution-based consensus-reaching model to address these issues. A convex clustering method is developed, combining the strengths of K-means and hierarchical clustering to enable adaptive subgroup formation and automatic determination of the optimal number of clusters. A new asymmetric implicit trust measure is developed by combining partnership dynamics with the Pearson Correlation Coefficient, providing a realistic representation of trust relationships. Furthermore, the model identifies leaders and members within each subgroup, quantifies their mutual influence through dynamic weights, and incorporates these dynamics into an improved Friedkin–Johnsen framework to allow for iterative preference adjustments and alignment toward consensus. The feasibility and validity of the proposed method are demonstrated through a case study and sensitivity analysis, highlighting its adaptability and effectiveness. Simulation experiments further validate the model, showing superior performance compared to existing methods.