AN ALGORITHMIC FRAMEWORK FOR DEVELOPMENT AND OPTIMIZATION OF FUZZY MODELS

Fuzzy models can be conveniently regarded as linguistic modelling stru ctures with well-defined functional blocks of input and output interfa ces along with a processing module. This paper examines the functions of these modules and specifies the relevant optimization tasks which a ccrue to them. Considering several levels of memorization completed wi thin fuzzy models (resulting in establishing short-, medium- and long- term memories), the corresponding learning policies are developed. Som e classes of optimization mechanisms are also discussed, with particul ar emphasis focused on their ability to handle multimodal problems. Th ese tools comprise both gradient-based techniques and selected methods of evolutionary optimization, in particular genetic algorithms. Illus trative numerical studies are also included.