Supplier Evaluation with MCDA: Ranking Stability and Weight Adjustment through Copeland and SITW

This study presents a comparative analysis of Multi-Criteria Decision Analysis (MCDA) methods applied to the supplier evaluation problem, with a focus on ranking stability and weight adjustment. Four prominent MCDA methods—FN-TOPSIS, FN-VIKOR, FN-MABAC, and SPOTIS—are used to generate individual supplier rankings based on a decision matrix comprising 34 alternatives and 12 evaluation criteria. To establish a unified reference point, a compromise ranking is constructed using the Copeland method. The main objective of the study is to determine how each method would need to adjust its internal weighting scheme to align with the compromise ranking. For this purpose, the Stochastic Identification of Weights (SITW) approach is employed. Fuzzy normalization using triangular fuzzy numbers is applied to stabilize the input data prior to analysis. Although fuzzy logic is used for normalization, all underlying data and evaluations are crisp. Ranking similarity is assessed using both the Weighted Spearman correlation coefficient and the Ranking Similarity Coefficient. Additionally, the Weights Similarity Coefficient 2 (WSC2) is used to evaluate the convergence of optimized weights. Results show strong alignment between FN-TOPSIS, FN-VIKOR, and FN-MABAC, both in rankings and re-identified weight vectors, while SPOTIS demonstrates a distinct internal logic and lower similarity with other methods. The proposed methodology offers a robust framework for understanding methodological behavior, enhancing decision transparency, and improving weight interpretability in complex supplier selection scenarios.