基于网格和排序的自适应遗传算法及其性能分析

Adaptive genetic algorithm based on grid and ranking and analysis of its property

提出了一种改进的自适应遗传算法.它使网格法思想在二进制编码表示个体的情况下得以实现,因此保证了初始种群内个体的遍历性和多样性;其遗传算子由个体在种群中的排序位置自适应地决定.该算法能维持种群内个体在各个运行阶段多样性,加快种群收敛速度,克服遗传算法早熟现象.几种典型的多峰函数优化结果证明该算法的有效性和实用性.

An improved adaptive genetic algorithm based on grid and ranking is presented. It makes the conception of grid accomplish successfully when individuals are figured bybinary code, thus guarantees the the popularity and diversity of individuals in the initial population. To sustain diversity of individuals in population during every operation phase efficiently, the probabilities of operator are also adaptively varied depending on the ranking value of individuals, Experimental results show that the improved algorithms can find the optimal individual quickly in solving the multimodal function optimization problems.