随机自适应差分进化算法

Self-adaptive Differential Evolution Algorithm with Random Mutation

针对差分进化算法存在早熟收敛且与理想最优值相差甚远等缺陷。随机自适应差分进化算法被提出,该算法采用随机选择策略的变异操作,再加小概率扰动;对变异因子和交叉概率进行自适应操作,以满足算法不同阶段的要求,其中交叉概率利用种群个体平均适应度值作对比,有利于充分利用种群信息。对几个标准函数进行测试并将该算法与其他4种方法进行比较,结果显示该算法的优化性能比其他方法好,具有较好的跳出局部最优的能力和收敛精度。

Aiming at the defects of differential evolution,such as the premature convergence and optimal value of differential evolution is far from the ideal optimal value.Self-adaptive differential evolution algorithm with random mutation was presented.Random choice strategy was adopted to execute mutation operation by the algorithm,which added to small-probability disturbance.To meet the requirements of different stages of the algorithm,the mutation factor and crossover rate performed the adaptive operation.The crossover rate was compared with the average of the fitness of individuals,which was beneficial to make full use of population information.Several standard functions were tested and the SRDE algorithm was compared with the other four methods.The results show that the optimization performance of SRDE algorithm is better than other methods,and it is better to jump out of local optimal ability and convergence precision.