A Hierarchical Fuzzy Systems Design Methods

As the application domain of fuzzy control expands from simple systems to more complex systems,a serious limitation of the standard fuzzy controller was discovered:"rule-explosion" problem,the rule base will quickly overload the memory and make the fuzzy controller difficult to implement. Therefore,many researches have been developed to deal with it. The hierarchical fuzzy system,proposed provides a way to deal with this problem. The hierarchical fuzzy system consists of a number of low-dimensional fuzzy systems connected in a hierarchical fashion. Design of a hierarchical fuzzy system has two aspects as usual:determination of structure and parameter optimization. Generally speaking,determination of structure has to solve two major problems:one is the linkage between subsystem;? and another is the arrangement of the input variables of each subsystem. The paper shows how to solve these problems. The method of clustering,genetic algorithm and Least Square Estimation are used in the paper. The arrangement of the input variables of each subsystem are finished by the use of the function ANFIS. The paper contains chapters as below:Chapter 1:Preface. Simply illustrating the background of the search topiccontrol and work of this paper done.Chapter 2:Hierarchical fuzzy system,Summarizing the structure of hierarchical fuzzy model with the most smallest rule numbers. Analyzing the parameters to be trained.Chapter 3:Structure determination of hierarchical fuzzy. Proposing a method of the arrangement of the input variables of each subsystem by use of ANFIS,using the definition of sensitivity. Introducing the application of subclust in designing hierarchical fuzzy system. Finding the clustering center with the function in MATLAB,obtaining the fuzzy rules,so the input space of each subsystem can be achieved. Chapter4:Parameter optimization of hierarchical fuzzy systems-the applicationof Least Square Estimation. Showing it is a linear function through analyzing the each subsystem,therefore the method of Least Square Estimation can be used.Chapter 5:Application of mixed Genetic Algorithm in hierarchical fuzzy system. Explaining detailedly algorithm flow and realizing program in hierarchical fuzzy model identification.Chapter6:System Simulation. We did CAD simulation on our algorithm ideology for nonlinear predictive systems,we gave the detail explanation in each step and some different examples for comparing. Chapter 7:Conclusion. Summarizing the work we have done in this paper and the questions still not resolved in our paper,we also gave the simple analysis on these questions.