Dynamic multi-species parallel differential evolution algorithm
Long Wen · Jisuanji yingyong yanjiu · 2012
Aiming at the problems of single population getting into premature convergence easily,this paper proposed a novel dynamic multi-population differential evolution algorithm.In this approach,it mtroduced the good point set method into the differential evolution(DE) initial step,which reinforces the stability and global exploration ability of the DE algorithm.Du-ring the evolution process,the proposed algorithm was based on individual fitness values,and divided the initial population into three sub-populations,and then they evolved with different DE algorithm by several trial vector generation strategies with a number of control parameter settings.It not only kept the independence of the sub-population and the superiority of the operators,but also not increased the complexity of algorithm.Tasted four classic benchmarks problems,and the experiment results show that the proposed algorithm is an effective method for different optimization problems.