大规模优化问题的差分进化策略对比实验研究  

A comparative experimental study on the differential evolution strategy for large scale optimization problems

作　　者：杨茂保[1] YANG Maobao(Jiujiang University, Jiujiang Jiangxi 332005, Chin)

机构地区：[1]九江学院,江西九江332005

出　　处：《智能计算机与应用》2016年第5期51-56,共6页Intelligent Computer and Applications

基　　金：九江学院校级科研课题(2014KJYB030);江西省高校省级教改项目(JXJG-14-17-10);江西省高等学校大学生创新创业计划项目(8891209)

摘　　要：差分进化算法(Differential Evolution,DE)有多种进化策略,并在求解各种优化问题时存在较大性能差异,求解大规模优化问题时不同策略间差异尤其明显。利用标准测试函数集对常用5种差分进化策略进行对比实验研究,分析这些策略的求解效果,总结求解大规模优化问题的不同差分进化策略适用性,为大规模优化问题的差分进化求解的策略选择提供帮助。Differential evolution algorithm（Differential Evolution,DE） has a variety of evolutionary strategies.And there exists a large performance differences between different strategies while solving multiple optimization problems. Especially, in view of large-scale optimization problem, this difference is particularly obvious. By using the standard test functions, 5 kinds of commonly used differential evolution strategy are compared.Based on the above,by analyzing the effect of these strategies, the paper summarizes the applicability of the differential evolution strategy of this research, which could provide the efficient help for the differential evolution strategy selection in sovling large scale optimization problems.

关 键 词：大规模 优化问题 差分进化 进化策略 对比实验 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]