论文部分内容阅读
基于空间交配遗传算法(GASM)采用空间交配遗传算子,有效克服早熟收敛问题,但缺少相关理论分析.文中采用马尔可夫链分析基于空间交配遗传算法的收敛性.证明采用最优个体保留机制的GASM,可收敛到全局最优解.同时证明在没有变异算子的情况下,GASM以概率1收敛到全局最优解.通过4个测试问题(其中3个为多峰值复杂问题)的对比实验,结果表明,GASM在求解多峰值复杂问题时,比采用最优个体保留机制的经典遗传算法,具有更好的收敛性.同时也与快速蜂群优化算法进行比较实验.
Spatial mating genetic algorithm (GASM) is used to overcome premature convergence problem effectively, but lack of relevant theoretical analysis.In this paper, Markov chain is used to analyze the convergence of spatial mating genetic algorithm, and the optimal individual retention mechanism GASM converges to the global optimal solution, and at the same time, GASM converges to the global optimal solution with no mutation operator at probability 1. By comparing the four test problems (three of which are multi-peak complex problems) The experimental results show that GASM has better convergence than the classical genetic algorithm with the optimal individual retention mechanism in solving multi-peak complex problems, and is also compared with the fast bee colony optimization algorithm.