摘要
在自适应遗传算法中交叉算子和变异算子随着其适应度变化自动改变其值,从而影响遗传进化的过程,但算法在进化初期对遗传操作的效果并不明显。本文针对离散变量的特征,通过计算个体间的离散程度,判断种群的进化程度,根据不同的进化时期自适应调整交叉概率和变异概率,使得种群的交叉和变异配合进行,有效地解决了离散变量在进化初期容易陷入局部寻优的问题。实验结果表明,算法经改进后,其全局收敛的可靠性增加并加快了收敛的速度。
The crossover operator and mutation operator in the adaptive genetic algorithm can adapt to change the value with the fitness value.The crossover operator and mutation operator will affect the evolution process,but the algorithm has little effective at the beginning of the evolution.This paper analyzes the characteristics of discrete variables,calculates discrete degree between individuals and judge the population evolution degree.The crossover probability and mutation probability work in cooperation,which can adjust their values in different evolutionary periods.The improved algorithm can resolve the problem of getting local optimization at the beginning of the evolution about discrete variables.The experiments show that the improved adaptive genetic algorithm increases the reliability of convergence to the global optimum and speeds up the convergence.
关键词
自适应
遗传算法
离散变量
进化系数
adaptive
genetic algorithm
discrete variables
evolution coefficient
