On the Efficiency of Multivariable TS Fuzzy Modeling

Studies on the efficiency of multivariable Takagi-Sugeno fuzzy modeling is described. From the interpretability viewpoint fuzzy modeling becomes complex especially in multidimensional cases due to decomposition of multidimensional fuzzy sets for each dimension. Even in one dimensional case, appropriate selection of fuzzy sets in number and shape is essential for satisfactory models. This general observation in mind, this research aims to make a study on the efficiency of fuzzy modeling taking two exemplary cases. Both cases are considered to be representing a dynamic system in two dimensional space. They are distinguished by the degree of nonlinearity they have. Both systems are modeled by fuzzy logic and the outcomes are comparatively studied to show explicitly how the modeling performance varies being dependent on the degree of nonlinearity involved.