摘要
为了提高遗传算法的收敛速度及局部搜索能力,设计了一种基于优良模式的局部搜索算子.同时对传统免疫算法中基于浓度的选择算子进行了改进,设计了一种基于适应度值和浓度的混合选择算子,从而有效的阻止了算法出现"早熟"现象.进一步给出了算法的步骤,并利用有限马尔可夫链证明了该算法的收敛性,最后通过对四个经典测试算法性能的函数的数字仿真,说明该算法对多峰值函数优化问题明显优于基本遗传算法.
To improve the convergence speed of genetic algorithm and local search capa- bilities, a local search operator based on excellent schema was designed. After Improving the selection operator of traditional immune algorithm based on the concentration, a mixed selection operator based on fitness value and the concentration was designed, which can effectively prevent the algorithm trapped into ＂premature＂ phenomenon. Further, The steps of the algorithm were given and the convergence of the algorithm was proved by Finite Markov chain. Digital simulation to four typical functions showed that the algorithm works effectively in multimodal function optimization problems than basic genetic algorithm.
出处
《数学的实践与认识》
CSCD
北大核心
2013年第11期177-184,共8页
Mathematics in Practice and Theory
基金
内蒙古工业大学科学研究项目(ZS201125)
关键词
局部搜索算子
模式
遗传算法
选择算子
local search operator
schema
genetic algorithm
selection operator
