A cluster-based evolutionary algorithm for multi-objective optimization

Abstract

In this paper a new evolutionary algorithm is described for multi-objective optimization. The new method handles non-linear objective functionsand constraints and supports the decision-maker with an estimation of thePareto set. This cluster-based method applies the Pareto-dominance principle. It approximates the Pareto set with the prototypes for each cluster and alternative prototypes as secondary population. The non-dominated set is continuously being up-dated: based on the Pareto ranking, the poorest clusters are regularly deleted, and the new ones are set. The method solves the usual test problems with a satisfactory level of accuracy. © Springer-Verlag 2001.