Supplier selection using a clustering method based on a new distance for interval type-2 fuzzy sets: A case study

• We introduce a suitable distance formula for interval type-2 fuzzy sets, which is compared with the Hamming distance, Euclidean distance, Hamming distance based on Hausdorff metric and signed distance.
• Based on the proposed distance formula, we propose a hierarchical clustering-based method to solve a supplier selection problem and find the proximity of the suppliers.
• To illustrate the applicability of the proposed method, two examples are illustrated. The first example is a case study in supplier selection problem. The next one is an example taken from the literature. Then, to test the hierarchical clustering-based method and compare the obtained results by two other methods, a comparative study using experimental analysis was designed.
• The novelties of our work are in both theory and application. We propose a new distance formula for interval type-2 fuzzy sets. Based on the proposed distance formula, we propose a new clustering method to solve a supplier selection problem and finally to illustrate the applicability of the proposed method, a case study is used.

Supplier selection is a decision-making process to identify and evaluate suppliers for making contracts. Here, we use interval type-2 fuzzy values to show the decision makers’ preferences and also introduce a new formula to compute the distance between two interval type-2 fuzzy sets. The performance of the proposed distance formula in comparison with the normalized Hamming, normalized Hamming based on the Hausdorff metric, normalized Euclidean and the signed distances is evaluated. The results show that the signed distance has the same trend as our method, but the other three methods are not appropriate for interval type-2 fuzzy sets. Using this approach, we propose a hierarchical clustering-based method to solve a supplier selection problem and find the proximity of the suppliers. To illustrate the applicability of the proposed method, first a case study of supplier selection problem with 8 criteria and 8 suppliers are illustrated and next, an example taken from the literature is worked through. Then, to test the hierarchical clustering-based method and compare with the obtained results by two other methods, a comparative study using experimental analysis is designed. The results show that while the proposed hierarchical clustering algorithm provides acceptable results, it is also conveniently appropriate for using interval type-2 fuzzy sets and obtaining proximity of suppliers.

Figure optionsDownload as PowerPoint slide