摘要
针对遗传算法局部搜索能力弱,求解精度不高的缺陷提出了一个基因定位算子.该算子的思路是进化一定代数(L)后通过对最优的若干个(N)染色体基因位从高位到低位逐次进行比较,如果当前的基因位都相同时便把该基因位确定下来,以后的交叉、变异操作都不让该基因位参考,随着算法的进行,染色体基因便从高位到底位逐渐地确定下来.基次,通过在基因定位过程中引入模拟退火思想和小生境技术等局部搜索能力的算法,提高该算子的全局优化能力.最后,通过几个非常容易陷入局部最优的测试函数测试表明几乎所有的峰值都得到了理论值.
<Abstrcat> A gene-orientation operator is proposed to solve the problem of weak ability and low precision in local searching in genetic algorithm.The operator makes a comparison among the several best (N) chromosomes from the high position to the low position after several generations (L).If all the genes on the current position happen to be the same,this gene is locked and not allowed to take part in the genetic operations such as crossover and mutation.In the process of algorithm,every gene of the chromosome will be determined gradually from the high position to the low position.Some algorithms with strong ability in local searching such as the Simulating Anneal and the Niche technology in the process of gene orientation are also used to improve the global optimizing ability of the operator.Finally,several test functions which are easy to fall into the local optimization are implemented to show that almost all the extremes reach their theoretical values.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2005年第3期491-494,共4页
Control Theory & Applications
基金
国家自然科学基金资助项目(60272099).
关键词
遗传算法
小生境技术
函数优化
基因定位
genetic algorithm
niche technology
function optimization
gene orientation
