A Cooperative Co-evolutionary Algorithm for Multi-objective Optimization

A new multi-objective optimization evolutionary algorithm based on the model of cooperative co-evolution is proposed in this paper. The algorithm incorporates the idea of elitism to motivate convergence, and adopts a novel form of collaboration between subpopulations, which improves its ability to keep diversity and avoids the difficult process of fitness assignment or non-dominance ranking in general multi-objective evolutionary algorithms so that the computational cost is greatly reduced. The proposed algorithm is compared with a well-known multi-objective evolutionary algorithm NSGA-Ⅱ on a suite of standard test functions using visual graphs and three quantitative metrics. Results indicate that this algorithm can search more effectively.