中国机械工程

• 智能制造 • 上一篇    下一篇

基于隔代映射算子的差分进化算法

符纯明1;姜;潮1;陈光宋2;吉磊2   

  1. 1.湖南大学汽车车身先进设计制造国家重点实验室,长沙,410082
    2.南京理工大学,南京,210094
  • 出版日期:2016-06-10 发布日期:2016-06-08
  • 基金资助:
    国家自然科学基金资助项目(11172096);教育部全国百篇优秀博士论文资助项目(201235);湖南省杰出青年基金资助项目(14JJ1016) 

Differential Evolution Algorithm with Intergeneration Projection Operator

Fu Chunming1;Jiang Chao1;Chen Guangsong2;Ji Lei2   

  1. 1.State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,Hunan University,Changsha,410082
    2.Nanjing University of Science and Technology,Nanjing,210094
  • Online:2016-06-10 Published:2016-06-08
  • Supported by:

摘要: 提出一种基于隔代映射算子的差分进化算法以求解优化问题,该方法在保证解的精度的同时具有较快的收敛速度。在经典的差分进化算法基础上,采用反向学习策略产生初始种群,并采用两种差分变异策略产生变异个体,以增加种群的多样性;利用隔代映射算子产生三个新个体替换当前进化种群中最差的三个个体,以实现精英策略提升算法的收敛性;为了保持种群的多样性和避免获得局部解,利用探测算子策略产生新个体加入进化种群。采用11个单峰、多峰测试函数和两个工程实例验证了该方法的有效性。

关键词: 差分进化算法; , 隔代映射算子; , 反向学习; , 探测算子

Abstract: A DE based on intergeneration projection operator with good optimum and fast convergence performance was proposed to solve optimization problems. The proposed method based on the classical differential evolution mainly included the following characteristics. Firstly, for improving the diversity of population, opposition learning was employed to generate initial population and two different strategies were randomly selected to generate new mutant individuals. Secondly, an intergeneration projection operator was designed to generate three offsprings to substitute for the three worst individuals into the next generation. Thirdly, the exploratory operator was introduced to generate the new individuals into the next generation for keeping the diversity of evolutionary population and avoiding to obtain local solution. Finally, the performances of IPDE algorithm were verified by the eleven single-and multi-modal benchmark tests and two practical engineering problems.

Key words: differential evolution(DE) algorithm, intergeneration projection(IP) operator, opposition learning, explorative operator

中图分类号: