System approximation via GA-based fuzzy model
You-Wei Teng, Wen-June Wang · Asia Pacific Conference on Circuits and Systems · 2003
This paper constructs a GA-based fuzzy model to approximate an unknown system without pre-set parameters. A real-coded GA and the least-squares method are applied to search for the antecedent and consequent parameters, respectively. Furthermore, a specific performance index is proposed to determine the optimal number of rules for the fuzzy model, and the adequate number of chromosomes and generations for the real-coded GA such that the approximation accuracy is satisfied.