摘要
通过对遗传算法初始种群与操作参数设定问题的研究,认为初始种群的分布状态与算子操作参数的选取直接关系遗传算法的全局收敛性与搜索效率,对初始种群与各操作参数进行合理设定是应用遗传算法进行寻优计算的重要问题.同时,遗传算法的初始种群必须科学地表征解空间的信息,操作参数也必须兼顾多样性与快速性相互协调设置.基于优化设计思想提出应用均匀设计方法同时确定遗传算法的初始种群及其他操作参数的方法.利用均匀设计的等价准则提出一种简化计算的近似获得均匀初始种群的方法,仿真实例验证了这种方法的可行性、有效性.
Based on the study on how to set the initial Fopulation and cperational parameters of operators, a conclusion can be drawn that distribution of the initial population and the selection of operational parameters of cperators directly concerns global convergence and searching effidency of genetic algorithm. The reasonable setting of initial population and operational parameters is an important problem in the application of genetic algorithm to performing optimization calculation. At the same time, the initial population of genetic algorithm must reflect the information on solution space scientifically, During the setting of opertional parameters attention must be paid to both diversity and fastness to coordinate them well. Based on optimization design theory, a method is proposed to establish initial population and cperational parameters simultaneously by uniform design. A simplified calculation method is thus proposed using equivalent principle of uniform design to obtain uniform initial population. Simulation results show that the method is feasible and effective.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第9期828-831,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60374003)
国家重点基础研究发展规划项目(2002CB312200)
关键词
遗传算法
初始种群
操作参数
收敛性
均匀设计
genetic algorithm
initial population
operational parameter
oonvergence
uniform design
作者简介
何大阔（1975-），男，辽宁沈阳人，东北大学副教授；
王福利（1957-），男，辽宁辽阳人，东北大学教授，博士生导师．
