NeuroFAST: on-line neuro-fuzzy ART-based structure and parameterlearning TSK model

摘要

NeuroFAST is an on-line fuzzy modeling learning algorithm,nfeaturing high function approximation accuracy and fast convergence. Itnis based on a first-order Takagi-Sugeno-Kang (TSK) model, where thenconsequence part of each fuzzy rule is a linear equation. Structurenidentification is performed by a fuzzy adaptive resonance theoryn(ART)-like mechanism, assisted by fuzzy rule splitting and addingnprocedures. The well known Δ rule continuously performs parameternidentification on both premise and consequence parameters. Simulationnresults indicate the potential of the algorithm. It is worth noting thatnNeuroFAST achieves a remarkable performance in the Box and Jenkins gasnfurnace process, outperforming all previous approaches compared