Research On The Comparison And Simulation Of Type-2 Fuzzy Logic Systems

The concept of Type-2 fuzzy set was introduced as an extension of the concept of an ordinary fuzzy set (henceforth called a Type-1 fuzzy set). Type-2 fuzzy logic system (FLS), which based on the Type-2 fuzzy sets, is a very powerful nonlinear signal processing technique that can handle the uncertainties. With the appearance this new type of FLS, the selection between Type-1 FLS and Type-2 FLS has become a new problem. Otherwise, the study of Type-2 FLSs can't be done without simulation. But, there is not convenient Type-2 FLSs simulation software. So these problems have to be solved.In the first part.of this paper, we discuss set theory on Type-2 fuzzy sets. Planar and three-dimensional figure of Type-2 fuzzy sets are presented. Then we show a example of the composition of the Type-2 fuzzy relations.Secondly, the basic structure of Type-2 FLS, especially the most popular interval Type-2 FLS, is presented. We also have discussed the comparison of Type-1 FLSs and Type-2 FLSs, then there are two conclusions: 1) When it comes to the forecast of chaos time series, we can using Lyapunov exponent as a parameter to chose which kind of FLS will be used to different case; 2) Type-2 FLSs which based on the Gauss-interval Type-2 fuzzy sets overcome Type-1 FLS much better when we got less information about the object.Thirdly, we designed simulation software which based on MATLAB for Gauss-interval Type-2 FLSs. By modularization designing, we carried out several module named selection of systems, training data input, parameter automatic setting, and training and application of Type-2 FLSs. We also provide a convenient help system for user. Then a complete example using this simulation software is presented.Finally, based on the former work, we using both Type-1 and Type-2 FLS to forecast stock market target which is extreme non-linear. As we expected, the performance of Type-2 FLS gain more advantage over Type-1 FLS.