Application Of ACS In Fuzzy Modeling

The theories and methods for fuzzy modeling are now the powerful tools to the modeling and control of the non-linear system which is complex and uncertain. So it's useful and valuable to research them.At first the factors which involve the performance of the fuzzy models are discussed. Based on these factors, three new methods are proposed to improving the performance of the fuzzy models.In order to improve the approximation accuracy, a kind of composite fuzzy model with error correction is proposed in ChapterⅢ.Then a new method for fuzzy parti- tion is proposed in ChapterⅤ.The new method is based on the ant-based clustering. It make the process of the input and output space fuzzy partition automatically. In chapterⅥ, the parameter identification is discussed. A kind of algorithm for continuo- us function optimization based on Ant Colony System is improved to suit for the parameter identification at first. Then a new kind of method for identification is proposed based on the improved algorithm. The new method can solve some complex identification problem that traditional methods can not solve.