一种基于GA的模糊控制规则优化新方法  

GA-based New Optimization Method for Fuzzy Control Rules

作　　者：邵克勇[1] 张鸿雁[1] 李飞[1] 谢维志 易江 

机构地区：[1]东北石油大学电气信息工程学院,黑龙江大庆163318 [2]中国石油辽河石化公司电工车间,辽宁盘锦124022

出　　处：《化工自动化及仪表》2011年第3期261-264,306,共5页Control and Instruments in Chemical Industry

基　　金：黑龙江省博士后科研启动基金资助项目(LBH-Q08159)

摘　　要：针对动态环境专家经验模糊控制规则效果欠佳的问题,提出一种改进的十进制基因编码自适应遗传算法,自动生成全局最优的控制规则。新算法引用稳态繁殖思想改进初始种群的生成方法,避免产生大量不合理个体影响进化进度,并引入动态变异率和一种新型的自适应变异算子调整种群的多样性,克服了"早熟"的现象。最后将获得的最优规则应用于设计模糊控制器。仿真结果表明,控制品质有较大的改善和提高。The expert experience-based fuzzy control rules produce poor effects in some unknown dynamic environments.An improved decimal-coded adaptive genetic algorithm and automatically-generated global optimal control rules were presented.The steady-state propagation idea,and a dynamic mutation rate and a new adaptive mutation operator were introduced into this algorithm to improve the way of generating initial population,and to accelerate evolution progress and adjust the population diversity as well as to end prematurity phenomenon.The simulation results show that the newly-designed fuzzy controller can perform well.

关 键 词：模糊控制规则 十进制 稳态繁殖 遗传算法 

分 类 号：TP11[自动化与计算机技术—控制理论与控制工程]