A Fuzzy Neural Network: Structure and Learning

Abstract

A promising approach to get both the benefits of neural networks and fuzzy logic systems and to solve their respective problems is to combine them into an integrated system such that we can bring the learning and computational power of neural networks into the fuzzy logic systems, and the representation and reasoning capabilities of fuzzy logic systems into the neural networks. For system modelling and control purposes their combination should provide an approach where structured knowledge of complex ill-defined systems is processed in a qualitative way, allowing reasoning and consideration of essential a priori information and performance criteria. Learning features should provide training procedures for synthesis, design, and implementation. Systems that combine neural network with fuzzy logic are called neurofuzzy systems.