Supporting multi-criteria decisions based on a hierarchical structure by taking advantage of acquired knowledge

In this paper, an approach to decision making is described. It combines a knowledge acquisition technique, a multi-attribute decision making technique, a validation technique and a machine learning algorithm. The suggested system is an extension of a previous decision support system based on the fuzzy repertory table technique. The aim is to increase its efficiency when dealing with a great amount of alternatives and criteria. The solution is based on a mechanism to divide the original problem into several simpler sub-problems. Moreover, a case study is presented to illustrate how the proposed system is used to design a Product Search Assistant. It will be integrated into a multi-agent architecture developed to give support to an e-marketplace.