Research on the Application of Convergence Ratio Parameter in Multi-Objective Evolutionary Algorithm
Libiao Zhang, Shuang Li, Chuankang Li, Chengqi Zhang, Bin Zhang, Meiyi Ge · IOP Conference Series Materials Science and Engineering · 2020
Abstract At present, evolutionary algorithms based on Pareto domination have been extensively studied. Sorting selection method is the most effective environment selection method in this kind of algorithm, which can effectively improve the convergence of the algorithm. But this method is prone to over-convergence of the population. Based on this, this paper proposes a convergence ratio parameter, using a sort selection method to screen a certain proportion of the solution, and the remaining places are selected using a binary game strategy. In this paper, by using this parameter in the KnEA algorithm and comparing it with the original algorithm, it is proved that the convergence ratio parameter can improve the diversity of evolutionary algorithms based on Pareto domination.