Further Developments and Tests of a Progressive Algorithm for Multiple Criteria Decision Making

P. Korhonen, H. Moskowitz, and J. Wallenius (1986) developed a progressive algorithm and the supporting theory for modeling and solving multiple criteria decision problems with discrete alternatives. A special feature of the algorithm is that it relaxes the usual assumption of a fixed set of available decision alternatives and complete knowledge of a decision maker's (DM's) preference structure or value function. The algorithm is based on progressively sampling the decision space, obtaining preference information from the DM, determining the likelihood of finding possibly/surely better alternatives, and based on this information, continuing the search or terminating it by making the final choice. We describe a computerized implementation and extensive computational tests of the algorithm, as well as some of our experiences in applying and field testing it in practice. We also consider several improvements and developments of the algorithm to further facilitate its use in practice.