摘要
为了保持进化过程中种群的多样性,提高算法的收敛速度,保护进化过程中的较优个体,对标准基因表达式编程(GEP)算法进行了改进,提出了一种基于适应度方差度量种群多样性的GEP算法(GEP based on population diversity measure by variance of individuals’fitness,DM-GEP)。该算法以个体适应度方差来度量种群多样性,设计了自适应变异算子,使得变异率随着种群多样性情况而变化,且同时兼顾了种群的稳定性以及进化过程中较优个体的保护。仿真结果表明,DM_GEP提高了收敛速度和精确度。
A GEP based on population diversity measure by variance of individuals＇ fitness （DM-GEP） is proposed aimed to keep the diversity of the group population, the convergence speed is improved, and the better individual is protected. Variance of indi viduals~ fitness is used to measure the group population diversity in DM-GEP, and the arithmetic designs an adaptive mutation operator, as the population diversity changes, the mutation probability changes, and gives the consideration to the stability of population and the protection of better individual. The simulation results show that DM- GEP improves convergence speed and accuracy.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第9期3094-3098,共5页
Computer Engineering and Design
基金
陕西省自然基金青年基金项目(2012JQ1019)
空军工程大学航空航天工程学院科研创新基金项目(XS1101021)
关键词
基因表达式编程
种群多样性
适应度方差
变异算子
自适应
gene expression programming
population diversity
fitness variance
mutation operator
self-adaptation
