MMOGA for Solving Multimodal Multiobjective Optimization Problems with Local Pareto Sets
Caitong Yue, Jing Liang, Ponnuthurai Nagaratnam Suganthan, Boyang Qu, Kunjie Yu, S. Liu
Abstract
Multiobjective optimization problems with multiple equivalent global Pareto solutions or with at least one local Pareto solution are called multimodal multiobjective optimization problems (MMOP). Most of the existing multimodal multiobjective algorithms can only find global Pareto solutions. However, the local Pareto solutions are of great significance when the global ones are impracticable. This paper proposes a Multimodal Multiobjective Genetic Algorithm (MMOGA) to find both global and local Pareto solutions. In MMOGA, only individuals in the same niche can mate and compete with each other, thus enabling the population to evolve in local areas. Experimental results show that the proposed algorithm can find both global and local Pareto sets of MMOPs.