一种融合Powell搜索法的差分进化算法

• 作者: 周永权 宁桂英 曹敦虔
• 来源期刊: 高等学校计算数学学报
• 学科: 工学
• 年，卷（期）: 2017,（3）
• 页码范围: 267-280
• 页数: 14页
• 分类号: TP183
• 语种: 中文

摘要

In the function optimization,in order to overcome the shortcomings of Differential Evolution Algorithm including precocious,slow convergence speed,easily falling into local optimum value,a Differential Evolution Algorithm Based on Powell Search Method is proposed.The Powell search method and the differential evolution algorithm are used alternately.The group optimized by the differential evolution algorithm has a certain probability to participate in the Powell search.By using the advantage of Powell search method including high precision,fast convergence speed and strong local search ability,the disadvantages of differential evolution algorithm is avoided effectively.Through the test of 10 classical functions,the simulation results show that the proposed algorithm outperforms the standard differential evolution algorithm in terms of convergence speed and convergence precision.