Optimization of hierarchical neural fuzzy models

Hierarchical fuzzy structures were introduced in previous work to deal with the dimensionality problem which is the main drawback to the application of neural networks and fuzzy models in the modeling and control of large-scale systems. In the present paper, the use of Radial Basis Function (RBF) networks connected in a hierarchical (cascade) fashion is investigated. The RBF networks are formulated as simplified fuzzy systems and the backpropagation equations for the optimization of the resulting hierarchical models are derived from this formulation. The optimization of the models using the conjugate gradient algorithm of Fletcher and Reeves is proposed and illustrated by means of a numerical example.