A Granular Computing Based Approach for Improving the Consistency of Intuitionistic Reciprocal Preference Relations

Abstract

Pairwise comparison between alternatives is the preference elicitation method most used in decision making. To model it, a mathematical model based on the concept of preference relation has been proposed. When preference relations are used, consistency, which is commonly related to the concept of transitivity, is a fundamental concern needing attention, especially when many alternatives are involved in the decision-making problem. Because inconsistent preferences could lead to not logical solutions, the study of consistency is essential in decision-making. Therefore, we develop in this study a new approach based on granular computing that improves the consistency when intuitionistic reciprocal preference relations are used to represent the preferences. This approach uses information granularity to develop the concept of granular intuitionistic reciprocal preference relations in which each entry is formed as an information granule instead of an exact numeric value. As well, it uses the multiplicative transitivity property to model the consistency associated with the intuitionistic reciprocal preference relations. Finally, an experimental example is shown to test and illustrate the performance of this approach.