An integration of AHP approach and Bayes classification algorithm in supplier selection

Abstract

Weight of criteria can only be changed through discussion or analysis of experts under traditional supplier selection method, which brings up two problems: (1) human effect problem and (2) incapability in timely decision making on updates of weights. In this paper, a new supplier selection method based on integration of Analytic Hierarchy Process, the Bayesian Classifier Algorithm and dynamic probabilities (AHP-BCA) is proposed. The method makes predictions with the probabilities of occurrences of criteria values based on historical records to avoid any human effects in decision making. It is also equipped with an instant self-update function to instantly update the probability values with new data, and be ready for next calculation. A simulation experiment is conducted to compare the performance of the proposed approach with a remarkable traditional approach in literature with historical data. Results show that the proposed approach can outperform the traditional one in achieving better selection results.