Robust modeling for nonlinear dynamic systems using a neurofuzzy approach with iterative optimization

Abstract

A neurofuzzy modeling approach for nonlinear dynamic systems is proposed in this paper. An iterative optimization approach for a class of neurofuzzy systems is developed, which integrates the model structure analysis and simplification, model parameter estimation, compatible cluster merging and redundant cluster deleting, performance evaluation for neurofuzzy models. The effectiveness of the proposed modeling approach is illustrated by the Mackey-Glass chaotic time series. The simulation studies show that the parsimonious neurofuzzy model is beneficial to the robustness of model. © Springer-Verlag Berlin Heidelberg 2005.