Hesitant fuzzy-based integrated multi-criteria group decision-making model for supplier selection

Abstract

In this study, a new integrated multi-criteria group decision-making (MCGDM) model under a hesitant fuzzy context is proposed to select reliable suppliers for the company. Different mutually conflicting criteria, human vagueness, and uncertainties are involved in this process. The selection process can be biased if the decision is made by a single expert instead of multiple experts. Classical fuzzy set theory (FST) is used to handle these issues. However, classical FST cannot completely handle uncertainty to some extent. On the contrary, the hesitant fuzzy set (HFS) is considered as one of the efficient and superior tools that can handle uncertainty completely. It can process and aggregate hesitant fuzzy information differently and capture their interrelationship. It can also handle uncertainty in the group decision-making process while experts have hesitation about several membership values for an element within a set. Due to its unique advantage, a new integrated MCGDM model is proposed which is consisted of the vague set, HFS, and weighted generalized hesitant fuzzy power geometric (WGHFPG) operator. The vague set is used to estimate the subjective weights of both the criteria and decision-makers; the HFS is used to consider multiple membership values in the decision-making process to handle uncertainty; and finally, the WGHFPG operator is used to aggregate the decision-matrices to get the final scores of the alternatives. The efficiency and practicability of the proposed model are illustrated by setting a numerical example. It is also benchmarked by other models including complex proportional assessment (COPRAS), combinative distance-based assessment (CODAS), and measurement alternatives and ranking according to compromise solution (MARCOS) by estimating Spearman’s ranking correlation coefficient. Finally, a sensitivity analysis is performed to test the robustness and stability of the introduced methodology. The result shows that the proposed model is more robust compared to other approaches.