改进遗传算法全局收敛性分析

传统的遗传算法大多数没有给出收敛性准则。一类新的改进的遗传算法被提出，该算法即考虑了优化问题的全局性要求——每一步构造一个新函数，而这往往却比局部最优理论和方法困难得多；同时通过对选择算子的改进，对遗传算法后期进化缓慢问题得到了有效控制，最后给出了算法的收敛性证明以及收敛性准则。实例证明该算法是有效的。

Global convergence analysis of improving genetic algorithm

The most traditional genetic algorithms didn't give a termination rule. A new kind of genetic algorithm is presented. In thisalgorithms, an algorithm for finding global minimization was proposed each phase must constructed a new function, which was moredifficult than local minimization. Meanwhile a selection operator was presented to make the place of the traditional one. It could preventthe latter slow evolution. The convergence of this algorithm is proved. A termination rule is given. The algorithm is efficiency provedwith some instances.