An Enhanced Decomposition-based Multimodal Multiobjective Algorithm Leveraging Neighborhood Information

The existence of multiple equivalent Pareto opti-mal sets in multimodal multiobjective optimization problems (MMOPs) poses new challenges to optimization algorithms. Despite much progress, decomposition-based algorithms have not yet fully realized their potential in solving MMOPs. This paper proposes an enhanced decomposition-based algorithm for MMOPs. The algorithm makes full use of the neighbor-hood information of the subproblems as well as the solutions in the decision space to generate high quality offspring, thus improving its convergence speed. In addition, a population pruning method is proposed to remove the less convergent solutions from the population so that computational re-sources can be used more efficiently. Experimental results show that the proposed algorithm significantly outperforms some existing algorithms in terms of overall performance in both the decision space and the objective space.