A Novel Multi-Objective Evolutionary Algorithm Based on External Dominated Clustering

Evolutionary algorithms (EAs) have wide applications in practice and many advantages over traditional methods in solving nonlinear and complex optimal problems. In this paper, we propose a novel clustering technique, in which the infeasible solutions are employed to divide the feasible solutions into several clusters. There is no more one infeasible individual in each cluster. A novel evolutionary algorithm based on this technique called ED-MOEA is proposed for dealing with constrained multi-objective problems. Simulation results on five test problems indicate the proposed algorithm is effective.