融入柯西扰动的改进差分进化算法及其应用

Modified Differential Evolution Algorithm Incorporating Cauchy Disturbance and its Application

为了避免差分进化算法提早收敛,提出一种融入柯西扰动的改进差分进化算法(CDMDE).使用双策略变异并在变异操作中加入柯西扰动和小概率扰动机制,提高算法的收敛精度;联合当前种群的中心解和最优解改进交叉策略,为算法提供良好的进化方向;自适应参数控制保留了优秀参数,有利于后续种群的进化;最优个体信息复制机制旨在挖掘种群中的优秀信息.通过优化19个测试函数,结果表明CDMDE算法与其他4种算法相比具有更高的收敛精度、更强的稳定性.并将该算法应用于2个电力系统经济调度问题,优化结果优于文献中所报道的结果.

To avoid differential evolution algorithm early convergence,a modified differential evolution algorithm incorporating cauchy disturbance（CDMDE）is presented. CDMDE operates the double-strategies mutation and adds the cauchy disturbance and disturbance with small probability mechanism in the mutation operation,which improves the convergence precision of the algorithm. CDMDE combines the central solution and optimal solution of the current population to improve the crossover strategy,which provides a promising evolution for the algorithm. The self-adaptive parameter control preserves the excellent parameters,which is beneficial for the evolution of the subsequent population. The optimal individual information replication mechanism is to exploit the excellent information in the population. By optimizing the 19 test functions,experimental results show that CDMDE has higher convergence precision and stronger stability than the other four kinds of algorithms. The proposed algorithm is applied to solve the two economic load dispatch problems. The solutions obtained by the proposed algorithm are better than those reported in the literatures.