Increasing the Accuracy of Calculations and Reducing Pairwise Comparisons Based on Clustering Decision Criteria and the Fuzzy Best-worst Method

In some cases, due to the ambiguity caused by qualitative judgments, the multiplicity of decision criteria, and the resulting increase in calculations, researchers use different methods to eliminate a number of criteria based on the employer's opinions. However, in real conditions, eliminating some seemingly unimportant criteria may ultimately lead to poor results in selecting the best option. This article seeks to address these shortcomings and present a new combined approach to increase the accuracy of calculations and reduce pairwise comparisons to determine the optimal weights of decision criteria. Therefore, first, the decision criteria are scored based on the combination of the decision maker's opinions and the Shannon entropy method and clustered using the K-means method. Then, the weighting process of the criteria in each cluster is performed separately using the best-worst fuzzy method. In order to ensure the accuracy of the results, several numerical examples are also provided. The results show that the proposed hybrid approach, in addition to preventing the elimination of some decision criteria, leads to increased accuracy in calculations and reduced pairwise comparisons between criteria compared to the best-worst method. In other words, the results of this approach, while requiring less comparative data, also provide more reliable answers to the decision maker.