遗传算法的扰动执行策略

Disturbance implementation strategy for genetic algorithm

从避免遗传算法陷入局部极优的角度,在分析遗传算子特性的基础上,提出了一种扰动执行策略,对交叉算子产生的新个体施加随机扰动,防止性能增益过小的个体模式在下一代中大量增长,有效保持了种群的多样性,从而使算法可以克服早收敛现象,实现全局搜索,并利用马氏链模型证明了算法的分布收敛性。经实验仿真验证,效果较好。

A disturbance implementation strategy is proposed for avoiding the local better solution in the global search with the genetic algorithm. Random disturbing is imposed on the new individuals generated by the crossover operator so that the number of schema of less increase is prevented from increasing drastically for keeping the diversity of population effectively. Therefore the genetic algorithm can overcome early convergency and realize all-round search. The distributed convergency of the algorithm is proved by Markov chain model,and simulation experiment shows that the effect is excellent.