摘要
分析了进化遗传算法的弊端。提出了一种基于实数编码和自适应变异率的改进遗传算法,将变异率定义为自上次进化以来未进化次数的函数,同时采取最优保存策略,有效地避免了采用二进制编码时计算精度与计算量之间的矛盾,克服了基本遗传算法因变异率选择不当引起的"早熟"现象及后期收敛速度慢的问题。计算结果表明,新方法成功地解决了进化遗传算法存在的问题,且计算效率较高。
The disadvantages of genetic algorithm based on real-coding evolution genetic algorithm were analyzed in this paper. An improved and adaptive mutation rate is presented, by defining the mutation rate as the function of the non-evolution time from the last evolution. The improved genetic algorithm effectively avoided the contradiction between the calculation precision of binary coding and calculation amount, eliminated the premature phenomena and overcame the problem of low convergence rate caused by the unsuitable mutation rate at the later convergence period by adopting the optimal that this new algorithm could solve the existing problems efficiency was high. preservation strategy. The calculation results indicated in evolution genetic algorithm and that the calculation efficiency was high.
出处
《重庆教育学院学报》
2008年第3期40-43,共4页
Journal of Chongqing College of Education
关键词
进化遗传算法
实数编码
自适应变异率
最优保存策略
早熟现象
evolution genetic algorithm
real-coding
adaptive mutation rate
optimal preservationstrategy
premature phenomena
作者简介
王璐（1980-），女，重庆市人，硕士，助教，主要研究方向：人工智能。
文武松（1982-），男，湖北公安人，硕士，助教，主要研究方向：智能控制及其应用。
