A Differential Evolution-Based Consistency Improvement Method in AHP With an Optimal Allocation of Information...

Area of Science:

• Decision Science
• Operations Research
• Artificial Intelligence

Background:

• The analytic hierarchy process (AHP) relies on reciprocal matrices from pairwise comparisons.
• Ensuring matrix consistency is crucial for reliable decision solutions in AHP.
• Existing methods often require modifications to improve matrix consistency.

Purpose of the Study:

• To present a new consistency improvement method for AHP using information granularity.
• To develop an optimal granularity model for maximal consistency in pairwise comparison matrices.
• To integrate decision-maker input into an interactive consistency improvement process.

Main Methods:

• Introducing information granularity to create granular rather than numeric comparison matrices.
• Developing an optimal granularity model to maximize consistency.
• Employing an interactive process with decision-maker involvement.
• Utilizing an adaptive differential evolution algorithm for matrix optimization.

Main Results:

• Demonstrated effectiveness of the proposed granular approach in enhancing matrix consistency.
• Achieved maximal consistency through optimal allocation of information granularity.
• Validated the method's performance via detailed experiments and comparative analysis.

Conclusions:

• The proposed granular consistency improvement method enhances the reliability of AHP decision solutions.
• Information granularity offers a powerful tool for optimizing reciprocal matrices.
• The interactive and adaptive approach provides a robust framework for decision analysis.