Group Consensus Formation Under Fuzzy Preferences: Addressing Hypocritical Trust and Irrational Confidence

With the rapid development of artificial intelligence (AI), group decision-making (GDM) has become a critical approach to ensuring the scientificity and fairness of decisions in complex systems. However, existing GDM methods often face limitations due to the presence of hypocritical trust and irrational confidence among decision-makers (DMs), which can significantly hinder the effective achievement of consensus. To address these challenges, this article proposes an innovative consensus reaching process that integrates fuzzy preference relations and advanced AI techniques to systematically identify and eliminate irrational factors in the decision-making process. The proposed method introduces a novel mechanism for quantifying the contribution degree of DMs based on game theory within the framework of fuzzy sociometric relations and effectively identifies and eliminates hypocritical trust relationships using geometric analysis. In addition, an innovative quartile classification method based on dynamic monitoring is designed to achieve real-time assessment and adjustment of the confidence level of DMs. By combining principles of calculus and geometry, an accurate calculation model of DM weights is constructed. The proposed method establishes a two-level consensus feedback mechanism integrating confidence level and trust degree, ultimately realizing the selection of the optimal alternative. Systematic case studies and comparative experimental analyses demonstrate the significant superiority of the proposed method in terms of consensus efficiency and decision-making quality. The key innovations of this research include the development of a new weight determination method based on calculus and geometry, the construction of a hypocritical trust identification mechanism using game theory and geometric analysis, and the establishment of an irrational confidence regulation system combining regret theory with dynamic monitoring. These contributions provide a robust theoretical framework and methodological support for addressing consensus challenges in complex GDM scenarios.