输入变量个数对T-S模糊建模精度的影响  

The Influence of the Number of Input Variables on the Descriptive Performance of T-S Fuzzy Model

作　　者：刘福才[1] 刘砚[1] 王树恩[1] 

机构地区：[1]燕山大学工业计算机控制工程河北省重点实验室,河北秦皇岛066004

出　　处：《模糊系统与数学》2013年第1期40-54,共15页Fuzzy Systems and Mathematics

基　　金：河北省自然科学基金资助项目(F2010001320)

摘　　要：输入变量个数会对模糊建模精度产生影响。对于一个实际的复杂系统,可测的或者需要考虑的输入变量非常多。是不是考虑的影响因素越多,即模糊系统的输入变量越多,则辨识的效果就越好呢?本文基于T-S模糊模型,分别采用对称三角形模糊划分和网格对角线法以及模糊聚类划分提取模糊规则,对Box-Jenkins煤气炉数据和Mackey-Glass混沌时间序列进行建模,得到了模糊模型训练性能指标和检验性能指标随输入变量个数增加时的变化趋势曲线,并给出了结论。The number of input variables in fuzzy modeling has a substantial influence on the accuracy of models. For an actual system which is fairly complex, there are many input variables that are measurable or need to be considered. Then, whether the more factor we consider, that is, the more input variables we use, the better identification results we get？ In this paper, based on T-S fuzzy model, the Box-Jenkins gas furnace data and the Mackey-Glass chaotic time-series are modeled by using the symmetrical triangular fuzzy partition and the fuzzy clustering algorithm. Through simulation, we get the changing tendency curve of the training and testing performance index in fuzzy models with increased numbers of input variables and give out the conclusion.

关 键 词：模糊辨识 输入变量个数 T—S模型 三角形划分 模糊聚类 

分 类 号：TP15[自动化与计算机技术—控制理论与控制工程]