Abstract
The modeling and identification of uncertain and nonlinear systems are important but challenging problems. Because of numerous advantages fuzzy models are often preferred to describe such systems. However, in many cases the generated models are very complex. Therefore, in this paper a combined method of fuzzy rules extraction, besides their simplification and optimization for creating a compact fuzzy rule base of Takagi-Sugeno (TS) models, is proposed that can be effectively used to represent complex systems. The initial fuzzy rule base is generated by using our proposed simple method of fuzzy rule extraction. The extracted rules are simplified using set-theory based simplification method and thereafter, simplified rules are further optimized using genetic algorithm. This, in turn, improves the accuracy of the model that was degraded during simplification. The novelty lies in the approach of fuzzy rule extraction method. The results are compared with those obtained by standard linear, neuro-fuzzy and advanced fuzzy clustering based identification tools.