基于双变异繁殖和HAD的多模态多目标差分算法OA

To improve the balance of algorithm performance in decision space and objective space,a multimodal multi-objective differential evolution algorithm based on double mutation reproduction and heterogeneity archive diversity(HAD)strategy(MMODE_DM_HAD)was proposed.The double mutation reproduction strategy was able to provide different mutation strategies for individuals of different grades to achieve a balance of population diversity and convergence.The HAD strategy archived the individuals separately according to their different properties and selected appropriate candidate solutions based on diversity criteria to achieve the maintenance of population diversity and improve the quality of solutions.Compared with 7 multimodal multi-objective optimization algorithms,MMODE_DM_HAD was validated on 14 test problems.Experimental results show that MMODE_DM_HAD ranks first in rHV values on all test problems,and it achieves the optimal IGDx values on 5 test problems and the suboptimal IGDx values on 5 test problems.While ensuring the performance of the objective space,MMODE_DM_HAD can fully explore the decision space and overall outperform the other comparative algorithms.