• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 人工智能与数据挖掘 • 上一篇    下一篇

基于精英种群策略的协同差分进化算法

马永杰,朱琳,田福泽   

  1. (西北师范大学物理与电子工程学院,甘肃 兰州730070)
  • 收稿日期:2017-04-21 修回日期:2018-01-23 出版日期:2019-02-25 发布日期:2019-02-25
  • 基金资助:

    国家自然科学基金(41461078)

A cooperative differential evolution
algorithm with elitist-population strategy

MA Yongjie,ZHU Lin,TIAN Fuze   

  1. (College of Physics and Electronic Engineering,Northwest Normal University,Lanzhou 730070,China)
  • Received:2017-04-21 Revised:2018-01-23 Online:2019-02-25 Published:2019-02-25

摘要:

针对差分进化算法在处理函数优化时存在的过早收敛和易陷入局部最优的问题,提出了一种基于精英种群策略的协同差分进化算法。在优化过程中,首先对种群进行适应度值评估和排序,提取前N个优秀个体组成精英种群,其余个体随机分为3个等大的子种群,每个子种群采取不同的进化策略,以此来保证种群的多样性;然后每隔一定代数,根据新的适应度值更新精英种群和其余3个子种群,这样可以有效地避免算法陷入局部最优;最后,将所提出的算法与4个先进的差分进化算法在CEC2014的30个标准测试函数上进行对比实验。实验结果表明,所提出的算法能够有效提高收敛速度,具有较高的收敛精度和较好的优化性能。

关键词: 差分进化, 函数优化, 精英种群, 协同进化, 参数自适应

Abstract:

To solve the problems of premature convergence and easy falling into local optimum in the differential evolution algorithm when dealing with function optimization problems of population diversity, we propose a cooperative differential evolution algorithm with elitistpopulation strategy (ESCDE). In the population optimization process, we firstly evaluate population fitness,  extract the first several excellent individuals as the elite population and divide the others randomly into three sub-populations of the same size. Each subpopulation adopts a different evolution strategy to guarantee the diversity of the population and improve the performance of the algorithm. And then, the elite population and the remaining three subpopulations are updated according to the new fitness value to avoid falling into local optimum. Finally, comparison are made between the proposed algorithm and four other advanced differential evolution algorithms on 30 standard test functions in CEC2014. Experimental results indicate that the proposed algorithm can improve convergence speed effectively with higher convergence precision and better optimization performance.

 

 

Key words: differential evolution, function optimization, elitist-population, cooperative evolution, parameter adaptive