多种群协同进化的差分进化算法

A differential evolution algorithm for multi-population coevolution

针对差分进化算法在优化过程中容易陷入局部最优和收敛精度不高的问题,提出一种多种群协同进化的差分进化算法。首先提出双序法用于种群划分:同时使用距离系数排序和适应度值排序将种群划分为三个子种群,将离全局最优个体远且适应度值优秀的个体划分出来,可以有效的避免陷入局部最优。其次对每个子种群采用不同的变异策略和控制参数,同时对整体表现一般的种群采用概率判定机制选择变异策略,以平衡全局探测和局部搜索。最后将所提算法在CEC2017测试集上进行实验仿真,实验结果表明,所提算法在收敛精度、跳出局部最优等方面均优于其他改进DE算法。

Aiming at the problems that differential evolution algorithm is easy to fall into local optimization and low convergence accuracy in the optimization process,a differential evolution algorithm for multi-population coevolution is proposed.Firstly,the double order method is proposed to divide the population.The population is divided into three sub populations by using distance coefficient ranking and fitness value ranking,and the individuals far from the global optimal individuals with excellent fitness value are divided out,which can effectively avoid falling into local optimization.Secondly,different mutation strategies and control parameters are used for each sub population,and the probability judgment mechanism is used to select the mutation strategy for the population with general overall performance,so as to balance global detection and local search.Finally,the proposed algorithm is simulated on the CEC2017 test set,and the experimental results show that the proposed algorithm is superior to other improved DE algorithms in terms of convergence accuracy and jumping out of local optimization.