摘要
提出了一种度量种群多样性的新指标 ,将其应用于交叉概率和变异概率两个参数的自适应调整 ,再将自适应的思想和并行计算的思想结合起来 ,提出了一种在个人计算机上实现的改进自适应遗传算法 -自适应伪并行遗传算法(APPGA) .对几种典型的多峰值函数求极值 ,结果表明 :该算法的全局搜索能力和收敛速度都远优于标准遗传算法 。
A new index which can reflect the diversity of population in genetic algorithms, is defined in this paper and used to adjust the probabilities of crossover and mutation. Based on the idea of self adaptation and parallel computing, an improved adaptive genetic algorithm is presented. The algorithm is called Adaptive Pseudo Parallel Genetic Algorithms (APPGA) because the parallel computing is realized on personal computer. The experimental results of optimizing several typical multimodal functions show that this algorithm has a great advantage of convergence property over standard genetic algorithms.
出处
《小型微型计算机系统》
CSCD
北大核心
2004年第7期1313-1316,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金 (5 9990 472 )资助
关键词
遗传算法
种群多样性
自适应
伪并行
收敛性能
genetic algorithms
diversity of population
self-adaptation
pseudo-parallel
convergence property