Eyad Y. Structural and parametrical optimization of digital controllers for fuzzy control systems

The thesis is devoted to the development of the methods and facilities for structural and parametrical optimization of digital fuzzy controllers (FC) for increasing of the fuzzy control systems efficiency in uncertainty. The method and software for input variables membership functions optimization based on the desired transient processes are developed; the gradient procedure for parametrical optimization of membership functions for output variables of both types FC (Sugeno and Mamdani) is improved; the method and facilities of FC structural optimization by choice of optimal number of rules and rules rating according to their influence on membership functions of fuzzy control signals are created; the set of Verilog-models for arithmetical-logic units and components of specialized table-based processors is developed. The efficiency of proposed methods and facilities is confirmed by results of FC structural-parametrical optimization for vacuum resistance furnace control.