ABSTRACT
In knowledge-based control the control signals are generated by an appropriate inference mechanism, and employing a control knowledge base which is typically expressed as a set of rules. In manual tuning of a control system a human expert tacitly uses a set of linguistic rules. Knowledge-based fuzzy tuning has its roots in a knowledge base of such rules. Computational requirement for knowledge-based tuning is addressed here in terms of the number of maximum and minimum operations needed to generate the tuning rule base and to obtain a tuning inference. The chapter presents concept of rule dissociation and the introduction of resolution relations to link the membership functions of the condition variables with those of the action variables form the basis of the analytical framework. In order to analyze the effect of fuzzy resolution on the accuracy suppose that the computational efficiency is not affected by the fuzzy resolution.
