Goal programming-based AHP criteria weight learning algorithm

An analytic hierarchy process (AHP) model was developed for vendor selection in a vertical E-commerce environment. The accuracy of the AHP-assisted decisions was improved with a goal-programming algorithm which adjusted the AHP criteria weights. The algorithm learns from the differences between a user's choices and the recommendations that the AHP model generates and can learn a decision maker's evaluation habits by averaging multiple learning results. Therefore, the algorithm greatly enhances the effectiveness of the AHP model. The algorithm was applied in a vendor selection system for a motorcycle industry E-commerce platform. Besides vendor selection problems, the algorithm is also applicable to other decision-making problems as well.