基于进化策略的模糊控制系统优化设计方法  

Optimal Design Approach for Fuzzy System Based on Evolutionary Strategy

作　　者：曹邦武[1] 姜长生[1] 文戎[2] 陈谋[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]西安交通大学电子与信息工程学院,西安710059

出　　处：《数据采集与处理》2002年第4期423-427,共5页Journal of Data Acquisition and Processing

基　　金：国家自然科学基金 (编号 :6 0 1 74 0 4 5 )资助项目

摘　　要：提出了一种改进的基于进化策略的算法 ,研究了模糊相似度、规则一致度的概念 ,将其量化从而影响模糊系统的进化过程 ,以根据样本数据设计出具有完备性、一致性和紧凑性的模糊系统。仿真结果表明 ,即使在学习样本较少的情况下 ,应用本文提出的方法 ,依然可以设计出具有完备性、一致性和紧凑性的模糊系统 ,而且该系统对初值不敏感 ,具备泛化能力。An improved algorithm based on evolutionary strategy is proposed. The concept of 'fuzzy similarity measure' and 'consistency of rule base' is studied, which is quantified to influence the evolutionary process in a fuzzy system. By this way a fuzzy system with completeness, consistency and compactness is designed from sample data. Simulational results show that a fuzzy system with completeness, consistency and compactness as well as the ability of generalization is acqurired by the design scheme. The fuzzy system is not sensitive to its initial status when lacking in enough sample data.

关 键 词：模糊控制系统 优化设计方法 进化策略 模糊逻辑 隶属函数 模糊控制器 模糊系统 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]