采用共生遗传算法的模糊控制器设计方法  

Fuzzy Controller Auto-Design Based on the Symbiotic Evolution Algorithm

作　　者：衣法臻[1] 胡恒章[1] 周荻[1] 

机构地区：[1]哈尔滨工业大学控制科学与工程系,黑龙江哈尔滨150001

出　　处：《系统工程与电子技术》2003年第6期750-753,共4页Systems Engineering and Electronics

摘　　要：为解决模糊控制器自动优化设计中对结构和参数的学习和寻优搜索缓慢 ,以及模糊规则基维数爆炸问题 ,采用共生进化遗传算法结合分级模糊建模的思想进行模糊控制器自动设计。提出了模糊规则分类形成子种群 ,多种群并行进化的遗传算法。用这种方法可以同时设计模糊控制器的结构和参数 ,极大地减小编码规模 ,提高搜索速度。对车 杆平衡系统的控制仿真实验表明 ,该方法寻优搜索速度快 ,设计出的控制器控制效果很好。To solve the problems of low learning speed and rule number explosion, symbiotic evolution combined with multistage fuzzy reasoning is used to design a fuzzy controller. The rule population is divided into several sub populations, and all these sub populations evolve in a parallel manner. With this approach, not only the structure and parameters of the fuzzy controller can be designed simultaneously, but also the coding of chromosome can be simplified and the solution searching speed is improved greatly. Simulations of cart pole control show that the performance of the controller designed by this method is fairly good.\;

关 键 词：模糊神经网络 遗传算法 系统辨识 模糊控制 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]