Consensus reaching process using personalized modification rules in large-scale group decision-making

Managing complex decision-making scenarios often hinges on the effectiveness of large-scale group decision-making (LSGDM). When confronted with a significant number of decision-makers (DMs) in LSGDM, each contributing unique backgrounds and perspectives, addressing the issues of reducing dimensionality and fostering consensus becomes a crucial aspect of the decision-making process. This paper addresses these challenges through several innovative approaches. First, we employ clustering methods to reduce the dimensionality of DMs. We introduce a novel fuzzy c-means clustering method that takes into account both the evaluation values and ranking of alternatives. This reduction in dimensionality serves to simplify the decision complexity and enhance the coherence of decision-related information among DMs placed within the same cluster. Once the clustering phase is complete, we propose a weight solution method for DMs within each group. This method combines the consensus level with the Spearman correlation coefficient of DMs, providing an effective means to determine the weights. Additionally, we introduce a weight solution method for each group based on the average consensus level and the number of DMs it contains. In the consensus reaching process (CRP), we implement a personalized modification rule. This rule takes into consideration the evolving consensus levels and the regret psychology exhibited by different DMs at different points in time. This dynamic approach significantly reduces both the cost and time required for consensus modifications. Finally, to validate the applicability of the proposed method, we apply it to a real-life case. Comprehensive qualitative and quantitative comparative analyses are conducted to evaluate the proposed method, along with a stability analysis of the parameters involved.