基于Fuzzy函数逼近的柔性模糊控制器论域压缩算法

Discourse of Universe Compression Algorithm of Flexible Fuzzy Logic Controller Based on Fuzzy Basis Function Approximation

针对模糊控制稳态控制精度不高的问题，提出了一种提高ＦＩＳ系统函数逼近精度的新方法——论域压缩算法．基于模糊数二元基元组的概念改进了一类Ｈｏｏｖｅｓ－Ｊｅｅｖｅｓ算法，实现对模糊数参数的寻优和输出论域的动态调整．构造基于ＳＯＷＡ类算子和ＢＡＤＤ解模糊的广义柔性模糊控制器作为仿真平台，对一过热汽温对象的控制给出仿真算例．结果表明，该方法可对控制达到任意精度上的逼近，从根本上解决Ｆｕｚｚｙ控制精度不高的问题．

For solving the lower control precision in fuzzy control, analysis of algorithm of FIS system was carried out. Fuzzy control is a kind of fuzzy basis function extension of the ideal control function from the viewpoint of mechanism of fuzzy inference system. It was also indicated that for the control targets the FIS implemented a local function approximation process. A novel method for increasing control precision (named discourse of universe compression algorithm) was developed based on the analysis. Based on the concept of two element basis of fuzzy number, an improved Hooves Jeeves pattern searching algorithm was designed to realize the parameter optimization of fuzzy number and dynamic tuning of discourse of universe of output value. A software simulation platform based on a universal flexible fuzzy logic controller using SOWA fuzzy operators and BADD defuzzification was constructed, on which the simulation of an overheated steam unit control was carried out. The results indicate that the method can achieve an arbitrary high approximation precision to ideal control and solve the problem of lower control precision of fuzzy control radically. The dynamic discourse of universe compression algorithm has a theoretical value for fuzzy control.