摘要
文中根据遗传算法理论分析了遗传编程中种群多样性对算法收敛特性的影响,提出了一种可行的种群多样性跟踪评测方法,同时提出了优选父代个体的改进方法。以求解旅行商问题为例,通过统计性实验数据验证了改进后的算法较采用同样局部优化的常规遗传算法具有更好的收敛速度和优化解,同时也对改进后算法的相关控制参数选择进行了实验分析,结论为改进算法能获得更好的收敛性能。
Narrates how population diversity affects the convergence property of genetic algorithms according the theory of genetic algorithrns in brief. Puts forward a feasible method tracking and evaluating the diversity of population, and brings forth the improving method - optimizing selection of parent individuals. Taking an example of traveling salesman problem (TSP) to validate above- mentioned idea, improving algorithms own better convergence property in comparison with canonical algorithms through the experimental statistic data. In the meanwhile, relative parameters of improving algorithms are analyzed also. The conclusion is that improving algorithms can get better convergence performance.
出处
《计算机技术与发展》
2006年第9期18-20,共3页
Computer Technology and Development
关键词
遗传算法
遗传编程
多样性
收敛特性
genetic algorithms
genetic programming
diversity
convergence properties