一种改进的自适应遗传算法

在对自适应遗传算法中选择、交叉、变异算子作用分析的基础上,提出一种新的自适应遗传算法,新算法基于实数编码机制,选择操作采用精英选择与轮盘赌相结合,变异和交叉操作采取根据适应度自适应地非线性调整变异和交叉概率的策略,同时提出进化的后期采取先变异后交叉的操作次序.仿真实验表明,新算法有效防止早熟,收敛速度更快,鲁棒性更好且拥有较强的寻优能力.

An Improved Adaptive Genetic Algorithm

Based on the analyzing of the function of selection,crossing and mutation,the new methods of adaptive genetic algorithm are discussed.Some improved measures are brought forward,for example,using real number for encoding,combining roulette wheel selection with elitist selection operator,using adaptive probability for crossing and mutation operator nonlinearly,simultaneously operating mutation before crossing in the evolution later period.The simulation shows that the improved genetic algorithm can prevent prematurity effectively and it has faseter convergence speed and higher roubustness with strong search efficiency.