Proposing an Ordered Clustering Based on the PROMETHEE Principles to Develop Purchasing Strategy in the Supply Chain

Purchasing portfolio models have received a great deal of attention in both academic and practice fields as suitable purchasing strategies. Purchasing portfolio applies as a diagnostic and prescriptive purchasing tool. The core purpose of this study is to introduce a quantified portfolio for developing purchasing strategies that are aligned with competitive priorities. The quantitative method of this study relies on data mining (ordered clustering) and MADM (Best Worst method) to classify purchased items with the aim of creating a strategic fit in the supply chain and developing purchasing strategies in accordance with the competitive priorities of organizations.

In portfolio models, the determination of the dimensions and the manner in which they are measured is important. In this study, firstly, the proper dimensions for commodity classification were introduced. These dimensions were competitive priority (as introduced in the literature review, including the costs, quality, speed, flexibility, and innovation), supply market analysis (an important dimension that should be considered in commodity classification analysis), and product features (that describe the characteristics of the commodities). Next, the proper criteria for each dimension were determined using the Delphi method. After that, the selected criteria using the Delphi method were weighted using the Best Worst method. In the following, purchasing items were classified using ordered clustering based on the PROMETHEE method. In this study, the clusters determined by PSO-K-means were ranked using the total unicriterion net flow of clusters in each dimension introduced in this study. Then, the preference profile was used to measure the preferential quality of each cluster on the different criteria in each dimension. The profile helps the purchasing managers with selecting the best proposed working methods for purchasing in each class of commodity by the preference profile. Finally, the proper working method and purchasing strategy were proposed for highly strategic commodities.

The approach and method presented in this study were implemented for 100 purchased items in a steel company. In this study, the company selected cost and quality as competitive priorities. The method construct bases on these competitive priorities and proper criteria for each dimension. The most appropriate purchasing methods for each class of high-strategic purchased items were presented taking into account the proposed methods and the opinion of experts. The strategies and working methods that were introduced in purchasing Chessboard were applied in this study.

The approach of this research helps purchasing and supply managers to have more and more accurate choices of purchasing methods for each category of purchasing items. By considering the profile of preferences in the ordered clustering method based on the PROMETHEE principles it helps them to improve supply management and supply chain performance. Also, the alignment of purchasing strategy with business strategy could improve competitiveness.