Modeling The Supplier Selection Process For Mechanical Parts

Official URL: https://risc01.sabanciuniv.edu/record=b3205827

As the complexity of outsourcing operations increases, the organizations’ procurement decisions get more complicated. Although there are various previous works about supplier selection for different industries, there exists no widely accepted universal model. This is due to the fact that organizations’ decisions are shaped differently according to their objectives and business environments. Moreover, even the same organization may need different selection criteria and methods for different products. In this study, a supplier selection model is developed for an organization that needs diverse mechanical parts which are produced by machining processes. We consider an organization that outsources the production operations of the mechanical part which are designed by the organization. Since the complexity of the operations is high and the parts are almost unique, manufacturability is a critical issue. Therefore, besides widely used criteria, the technical competencies of the suppliers are considered as well. The scope of this study is to determine the shortlist of the suppliers to be asked for a quote, rather than evaluation of the quotes. The motivation here is that although having more quotes is beneficial for the organization, it is a costly operation. Therefore, determining the ideal set of suppliers for a product is valuable. In this study, an ideal criteria set is determined upon interviews with professionals. Then, importance levels (i.e., weights) of the criteria are determined by the procurement staff. As the next step, evaluation procedures are defined for criteria. For the evaluation, extensions such as distributed decision-making and decision-making with missing data are used. Afterward, a fitting score is found for supplier and mechanical part pairs, on which supplier recommendations are made. For fitting scores, besides the cost and benefit perspectives on the value functions, the target value approach is utilized.