Abstract: The convergence speed of Genetic Algorithm(GA) and the quality of problem result are the main inconsistency which affects the performance of GA. This paper proposes the control strategies of improved GA, which are the good subspace operator, the variable crossover probability and variable mutation probability operator and the variable dimension subspace operator. Experimental results show that the convergence speed of this algorithm is so fast to improve GA.

Key words: Genetic Algorithm(GA), TSP problem, control strategy

摘要： 传统的遗传算法收敛速度与问题解的质量是影响算法寻优性能的一对矛盾。该文提出一种新的遗传算法的控制策略——精英子空间算子、变交叉概率Pc和变异概率Pm算子和变维子空间算子。实例计算表明该算法收敛速度快，可以进一步改善遗传算法的性能。

关键词: 巡回施行商问题, 控制策略

CLC Number: 

• TP18