A PSO-BASED METHOD FOR EXTRACTING FUZZY RULES DIRECTLY FROM NUMERICAL DATA

In this paper, a PSO-based method is proposed for automatically constructing a fuzzy system with an appropriate number of rules to approach the identified system. In the PSO-based method, each individual in the population is constructed to determine the number of fuzzy rules and the premise part of the fuzzy system, and then the recursive least-squares method is used to determine the consequent part of the fuzzy system constructed by the corresponding individual. Consequently, an individual corresponds to a fuzzy system. Subsequently, a fitness function is defined to guide the searching procedure to select an appropriate fuzzy system with the desired performance. Finally, two identification problems of nonlinear systems are utilized to illustrate the effectiveness of the proposed method for fuzzy modeling.