遗传算法的收敛性分析

Convergence analysis of genetic algorithms

用随机过程论中的马尔克夫链理论研究了几种遗传算法的收敛性 .提出了 6个引理 ,3个定理和 3个推论 ,证明了最优保存遗传算法和作者提出的 2种新型遗传算法 :模拟生物种族进化的遗传算法 ,带罗盘算法的GA是全局收敛的 ,而简单遗传算法不是全局收敛的 .

The convergences of several genetic algorithms are studied by using Markovian chain theory in stochastic processes.Six lemmas, three theorems and three corollaries are proposed.It is proved that optimum maintaining simple genetic algorithm and two new genetic algorithms put forword in this paper are global convergent,but simple genetic algorithm is not the case.