具有广义线性隶属函数的典型模糊系统的通用逼近性  

Universal approximation of typical fuzzy systems with generalized linear membership function

作　　者：黄卫华[1] 方康玲[1] 章政[1] 陈士杰 

机构地区：[1]武汉科技大学冶金自动化与检测技术教育部工程研究中心,武汉430081 [2]中冶连铸技术工程股份有限公司,武汉430073

出　　处：《计算机应用研究》2010年第4期1263-1265,1269,共4页Application Research of Computers

基　　金：国家"863"计划资助项目(2007AA04Z100);湖北省自然科学基金资助项目(2007ABA280)

摘　　要：设计了一种将三角形和梯形隶属函数作为特例的广义线性隶属函数,推导了输入采用广义线性隶属函数的典型Mamdani模糊系统的解析结构,证明了典型模糊系统是单调、递减的有界连续函数;在此基础上证明了该类模糊系统能以任意精度逼近任意连续实函数,最后仿真实例证明了本设计的有效性。This paper designed a kind of membership function named generalized linear membership function,and triangle and trapezoid shaped membership function were as its special cases.Deduced the analytical structure of the Mamdani fuzzy system whose input sets were with generalized linear membership function and manifested that the output of the fuzzy systems were monotonical,decreasing,continuous and bounded.On the basis of the analytical structure,it was proven that the fuzzy system could approximate any real continuous function.Finally,gave some simulations to manifest the effect of the design.

关 键 词：广义线性隶属函数 模糊控制 通用逼近性 解析结构 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]