A hybrid approach to supplier performance evaluation using artificial neural network: a case study in automobile industry

For many years, purchasing and supplier performance evaluation have been discussed in both academic and industrial circles to improve buyer-supplier relationship. In this study, a novel model is presented to evaluate supplier performance according to different purchasing classes. In the proposed method, clustering analysis is applied to develop purchasing portfolio model using available data in the organizational Information System. This method helps purchasing managers and analyzers to reduce model development time and to classify numerous purchasing items in a portfolio matrix. In this paper, Neural Networks are used to develop a purchasing classification model capable of classifying purchasing items according to different features. Moreover, a new supplier evaluation model based on different purchasing classes is developed using Neural Networks. The proposed hybrid method to develop purchasing portfolio and supplier evaluation is applicable in large scale manufacturing organizations which need to manage numerous purchasing items. The proposed model is implemented in an automaker purchasing department with a relatively vast supply chain and the results are presented.