A Dynamic Multi-Objective Differential Evolution Based on Hierarchical Prediction Strategy

In order to effectively improve the convergence of the dynamic multi-objective evolutionary algorithm, this paper proposes a dynamic multi-objective differential evolution based on the hierarchical prediction strategy, which is combine into NSGA-II, and noted as (HPS-NSGA-II/DE). Firstly, the degree of environmental change is calculated by the correlation coefficient, and then the prediction strategy is activated adaptively according to the degree of change. In addition, the mutation operation of differential evolution is improved based on the local optimal solution, which can improve the searching ability of differential evolution. Finally, the performance of the proposed algorithm is verified by experimental verification on 10 standard test functions.