摘要
为了改善传统自适应遗传算法的收敛速度以及局部收敛问题,根据种群适应度的集中程度,以种群的最大适应度、最小适应度以及适应度平均值这3个变量为基础,设计了改进的自适应交叉概率和变异概率来调整整个种群的交叉概率和变异概率,提出了一种基于种群适应度集中程度的改进自适应遗传算法.将该算法应用于函数优化中,仿真结果验证了其具有"快速收敛"的特点,且在很大程度上可避免遗传算法的早熟现象.
To speed up convergence rates and resolve local convergence issues in traditional adaptive genetic algorithms, an improved adaptive genetic algorithm was developed. According to the concentrating degree of fitness of the populations, a kind of adaptive crossover probability and mutation probability were de- signed in terms of three variables of maximal fitness, minimal fitness and average fitness of the popula tions, whereby the crossover probabilities and mutation probabilities of the＇ whole populations could be ad- justed. Based on this, an improved adaptive genetic algorithm was developed. Simulation results prove that the new adaptive algorithm can converge faster than the unimproved algorithm and is highly effective at avoiding the premature convergence of the adaptive genetic algorithm.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2007年第8期875-879,共5页
Journal of Harbin Engineering University
基金
黑龙江省博士后基金资助项目(AUGA41000542)
黑龙江省自然科学基金资助项目(F2004-19)
关键词
自适应遗传算法
交叉概率
变异概率
函数优化
全局收敛
adaptive genetic algorithm
crossover probability
mutation probability
function optimization
global convergence
