Regret-Theoretic Multiattribute Decision-Making Model Using Three-Way Framework in Multiscale Information Systems.

Area of Science:

• Decision Sciences
• Information Systems
• Computational Intelligence

Background:

• Existing multiattribute decision-making (MADM) methods on multiscale information systems (MSISs) often rely on utility theory, leading to rigid classifications and difficulty in reflecting decision-maker risk attitudes.
• These limitations can result in misclassification and fail to capture the nuanced psychological behaviors of decision-makers.

Purpose of the Study:

• To propose a novel wide three-way decision (3WD) model for MSIS that integrates 3WD theory with regret theory.
• To overcome the shortcomings of existing MADM methods by improving classification accuracy and incorporating decision-maker risk preferences.

Main Methods:

• Constructing an outranking relation on the MSIS using the regret-rejoicing index based on regret theory.
• Classifying objects into three domains using a clustering method.
• Calculating the ranking of objects within each domain using the relative closeness coefficient.

Main Results:

• The proposed 3WD model effectively addresses the limitations of traditional utility-based MADM methods in MSIS.
• Experimental simulations and comparative analyses demonstrate the model's effectiveness, superiority, and stability.
• The model accurately classifies objects and better reflects decision-maker behavior and risk attitudes.

Conclusions:

• The integrated 3WD and regret theory model offers a more robust and nuanced approach to decision-making in MSIS.
• This novel model enhances classification accuracy and decision-maker representation compared to existing methods.
• The study validates the proposed model's practical applicability and effectiveness through empirical testing.