针对传统的简单遗传算法的缺陷,提出了改进的具有成熟前收敛判断的自适应遗传算法.用群体熵值和均方差来预报成熟前收敛的发生.当成熟前收敛发生时,提出以群体中的最优个体为基础,在其一定大小领域内随机产生若干个体,取代原种群中的部分个体,其中更新的个体数占群体中个体总数的30％～40％,领域大小与目标函数极值点分布有关.仿真实验证明,算法的收敛速度和全局收敛概率都有显著的提高. Aiming at the defects of the traditional simple genetic algorithm, an improved adaptive genetic algorithm with pre-mature convergence judgment is proposed, and the pre-mature convergence is forecasted by the population entropy and mean square error. When pre-mature convergence occurs, Based on the optimal individuals in a certain size of the area of ​​a number of randomly generated individuals to replace some of the original population of individuals, of which the updated number of individuals in the population accounted for 30% to 40% of the total population, the size of the field and the extreme value of the objective function Point distribution.The simulation results show that the convergence speed and the global convergence probability of the algorithm are significantly improved.