摘要
针对差分演化算法与进化策略算法中所存在的不足,将模拟退火算子引入到差分演化算法的变异操作中,这样有助于在进化前期进行全局搜索,后期进行局部搜索;在标准进化策略的基础上,加入差分变异操作,提出了一种新的差分进化策略双重变异算法。通过测试算例可看出,该方法在多峰值函数优化问题中,具有求解精度较高,收敛速度较快等特点。
Against to the finite about differential evolution algorithm and evolution strategy,this paper brings the simulated evolutionary operator into the differential evolution algorithm,it can help enhance global search in prophase and partial search at later period when evolving.Based on the normal evolution strategy adding the differential mutation operator in it,a new Bi-mutation differential evolution strategy algorithm is proposed.From the following examples,it can be seen that the result of the multiple hump function is very accurate and the convergence speed is fast.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第32期41-44,共4页
Computer Engineering and Applications
基金
国家自然科学基金No.60461001
广西自然科学基金No.0542048
广西研究生教育创新计划资助项目(No.2007106080701M18)~~
关键词
退火因子
差分演化算法
进化策略
双重变异
anneal operator
differential evolution algorithm
evolution strategy
Bi-mutation
