摘要
针对标准遗传算法中交叉算子容易破坏定义长度较长的优良模式的弱点 ,提出了一种动态排序编码方法 ,以提高交叉算子的效率 .首先对当前代种群确定基因权重 ,然后根据基因权重对基因位置进行动态排序 ,使优良基因变得集中 ,从而克服了标准遗传算法中交叉算子的弱点 .为了避免陷入局部最优 ,对变异算子进行了改进 .最后做了大量实验 ,并根据实验结果对该方法进行了评述 .
Crossovers, in the standard Genetic Algorithm(SGA), tend to destroy excellent schemata whose defining lengths are comparatively long. Therefore, we propose the method of dynamic ranking encoding, to improve the performance of crossover. This method, firstly, determines gene weights in the chromosome for the current population, and then ranks gene loci dynamically by gene weights to make excellent genes concentrated. Thus, the shortcoming of crossovers in SGA is overcome. Moreover, we improve the mutation to avoid the remaining local optima of GAs. At last, many experiments are done, and according to the results of these experiments, we evaluate this method.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2003年第11期14-19,共6页
Systems Engineering-Theory & Practice
基金
国家自然科学基金 (70 1 71 0 0 2
699740 2 6)
关键词
遗传算法
基因权重
动态排序编码
genetic algorithms
gene weights
dynamic ranking encoding
