An Improved Progressively Interactive Evolutionary Multi-objective Optimization Algorithm with a Fixed Budget of Decision Maker Calls

This paper presents a preference-based method to handle problems with a large number of objectives. With an increase in number of objectives the complexity of the problem rises exponentially and it becomes difficult for evolutionary multiobjective techniques to produce the entire front. In this paper an evolutionary multi-objective procedure is hybridized with preference information from the decision maker during the intermediate runs of the algorithm. The preference information from the decision maker helps in guiding the algorithm towards the most preferred point on the front. The methodology is an improvement of an earlier progressively interactive approach which uses implicitly defined value functions. The proposed approach offers multiple advantages over the previous technique with the most important advantage being optimizing the problem in a fixed budget of decision maker calls. In the previous method, there was no control over the number of decision maker calls required to optimize the problem. The suggested approach works by constructing polyhedral cones which are used to modify the 1Also, Department of Information and Service Economy, Aalto University School of Economics, PO Box 21220, 00076 Aalto, Helsinki, Finland KanGAL Report Number: 20110014 domination principle and conduct a focussed search. The methodology has been tested, and comparison has been performed with the earlier approach, on two to five objectives unconstrained as well as constrained test problems.