基于量子粒子群的进化聚类设计模糊控制器  

Design of Fuzzy Controller with Evolutionary Clustering Based on Quantum-behaved Particle Swarm Optimization

作　　者：薛福珍[1] 唐雷[1] 刘旭[1] 

机构地区：[1]中国科学技术大学自动化系,安徽合肥230027

出　　处：《控制工程》2009年第5期598-601,605,共5页Control Engineering of China

基　　金：安徽省发改高技基金资助项目(2005661)

摘　　要：针对模糊控制器设计困难的问题,提出用基于量子粒子群的进化聚类方法从数据中分析提取模糊规则以设计模糊控制器,并提出新型的进化编码方案以解决需要预先确定聚类数目的问题。据此方法建立了可视化CAD平台,然后进行工程化实现,用所设计的模糊控制器对水箱液位进行控制。运行结果表明,较之传统聚类方法设计模糊控制器,设计的模糊控制器具有更好的动态和稳态性能,从而证明了该设计方法及相应CAD平台的有效性和实用性。To the difficulties in designing of fuzzy controllers, a method is proposed based on quantum-behaved particle swarm optimization. The fuzzy controller is designed by implicitly extracting fuzzy rules from data using evolutionary clustering method. A coding scheme is presented to solve the problem of setting the cluster number beforehand. A visual CAD platform is constructed, and the designed fuzzy controller is used on the platform to control the water tank level. Compared with fuzzy controller designed using traditional clustering method, the design result shows the dynamic and steady performance and the effectiveness of the proposed method and the practicability of the CAD platform.

关 键 词：量子粒子群 聚类 模糊控制器 CAD平台 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]