Fuzzy sequential three-way multi-scale attribute decision-making method based on individual risk attitude and multi classification features

Multi-attribute decision-making research is of great significance for solving macro problems. However, the existing multi-attribute decision-making methods face two problems: one is how to comprehensively consider the impact of irrational behavior on the decision-making results; the other is how to make intelligent decisions on the evaluation information of “multi-level, multi-classification, multi-perspective”. To address the above two issues, this paper establishes a fuzzy sequential three-way multi-scale attribute decision-making method based on individual risk attitudes and multi-classification features. First, we construct multiple attribute combinations from the inconsistent multi-scale attribute set and weight and aggregate them into comprehensive decision attributes, thereby transforming them from multi-scale to multi-view. Next, we identify multiple attribute clusters through hierarchical clustering and create a class-cluster dependency definition to determine the sequential set using a heuristic algorithm. We then propose a specific sequential three-way decision model within the framework of granular computing, tailored to the characteristics of the evaluation information. For object ranking, we pre-rank the objects based on regret theory and develop two methods to determine category weights based on the classification results obtained from the three-way decision. The stability and effectiveness of the proposed method are verified through corresponding experiments and comparative analysis of real cases.