一种基于方差的自适应遗传算法

An Adaptive Genetic Algorithm Based on Square Error

为了解决遗传算法的收敛速度和全局收敛性之间的矛盾,本文提出了一种改进的自适应遗传算法Adaptive GA Based on Square Error(SEAGA)。在原自适应遗传算法Adaptive GA(AGA)的基础上提出用适应度方差函数来监控种群的进化情况并据此自动调整算法的交叉率和变异率的思想。通过用此算法对测试函数进行计算,并与SGA,AGA的结果进行比较,可以看出本算法在收敛速度和全局搜索性上优于其它同类算法。

An improved adaptive genetic algorithm based on square error-SEAGA is proposed to solve the contradiction between the speed of convergence and the ability of global convergence in genetic algorithms. This algorithm uses square error function to monitor the evolvement of the population and tune the probability of crossover and mutation automatically. Compared with the Simple CA and the Adaptive CA, the SEAGA is better in the speed of convergence and the ability of global search.