摘要
文中对遗传算法中常用的交叉算子的搜索能力进行了讨论,综合考虑了交叉算子对染色体的双重作用,给出了染色体在交叉算子作用下数量和分布范围的变化规律,得到了模式变化的极限性能,证明了交叉算子在演化代数增加的时候能够使模式内部各基因趋于独立,并且只要组成模式的各个基因都存在,则该模式一定能够被搜索到,此时模式的极限概率等于组成该模式各基因的初始概率(也就是基因的极限概率)的乘积,并且与模式的定义长度无关,说明了交叉算子使群体分布扩散的特性。
In this paper, the search capability of common crossover operator is discussed. By studying the dual function of the crossover operator, the rule of the amount and range of the chromosome are obtained. It was proved that the genes would be independent. It was also proved that the limit probability of the schemate equals the product of the initial probability of the genes and had nothing to do with the definition length of the schemate. By this study the performance of the genetic algorithm can be improved.
出处
《计算机研究与发展》
EI
CSCD
北大核心
1999年第11期1317-1322,共6页
Journal of Computer Research and Development
基金
国家自然科学基金
上海交通大学博士启动基金
关键词
遗传算法
交叉算子
搜索能力
genetic algorithms, schemate, crossover operator, search capability