A Multimodal Multi-Objective Evolutionary Particle Swarm Algorithm Based on Grey Prediction Selection Strategy

Abstract

In view of the fact that the diversity of the objective space of multimodal multi-objective optimization will decrease with the increase of the diversity of the decision space, and it is unable to identify all pareto optimal solution sets, this paper proposes an evolutionary particle swarm optimization algorithm based on the gray prediction selection strategy to solve this problem. First, the historical optimal individuals and the neighborhood optimal individuals are selected by using multivariable gray prediction to improve the diversity of the target space; Secondly, the index based ring topology is used to update the neighborhood optimal archive, induce a stable ecological niche, and identify more pareto optimal solution sets; Finally, the non dominated special crowding distance is used as a density measure in the decision space and objective space to identify more pareto optimal solutions. In the experimental stage, 11 MMO test functions are selected and compared with four classic multimodal optimization algorithms. The results show that the proposed algorithm can identify all pareto optimal solution sets in the decision space and maintain good convergence and diversity in the objective space.