Computer aided tuning and validation of fuzzy systems

A set of tools to support the design of fuzzy controllers is illustrated. The tools, based upon dedicated optimization algorithms, allow the tuning and the validation of the fuzzy models. A modified Monte Carlo method has been developed to define stochastic algorithms to complement the deterministic approach, because of the peculiar structure of the systems addressed. The authors show that the design tools are suitable for giving a consistent answer to the tuning needs of fuzzy control systems. During experiments the proposed algorithms have successfully lead to model optimization allowing the fulfilment of design requirements.<
>