A novel three-way decision utility model considering decision-maker’s attribute preferences

In the study of Multi-Attribute Group Decision Making (MAGDM), determining the weights of decision-makers is a core aspect, as it directly affects the fairness and rationality of decision outcomes. The varying risk preferences of decision-makers also significantly impact the final decision. The Three-way Decision (TWD) utility model, a powerful tool for dealing with uncertainty and fuzziness, has shown its unique advantages in solving classification problems, primarily considering conditional probability and utility functions. This paper first proposes an optimization model for updating decision-makers’ weights based on their objective weights and subjective information. The Newton iteration method is employed to find the optimal solution iteratively. Subsequently, a method for determining conditional attributes using grey relational analysis (GRA) is introduced. And a new utility function is developed to reflect the individual risk attitudes and attribute preferences of decision-makers. Based on these developments, the paper constructs a new TWD utility model tailored for MAGDM. The paper demonstrates the model’s effectiveness and superiority through a case study involving real-world data. By comparing it with other methods, the model’s advantages are highlighted. A sensitivity analysis of the model’s parameters is conducted to test its robustness to changes, and the model is applied to large datasets, validating its potential and applicability in practical applications.