Optimization of fuzzy controllers by exhaustiv, trial & error and genetic methods. Medical applications
Lucian Vasile Boiculese, Horia-Nicolai Teodorescu, Gabriel Dimitriu · DOAJ (DOAJ: Directory of Open Access Journals) · 1996
The membership functions and rules optimization of a fuzzy controller are focused in this paper. Four optimization methods are addressed: (i) exhaustive method, (ii) trial and error, (iii) genetic algorithm, (iv) a combination of (ii) and (iii). The methods were tested in an application for anesthesia control, using a fuzzy controller with two inputs and one output, implemented on a PC. The results obtained by the classic method, by the trial-and-error (T.E.), by genetic algorithms (G.A.) and by a mixture of G.A. & T.E. are contrasted. The number and type of the membership functions (m.f.), are also taken in account. It is shown that a combination of G.A. & T.E. gives best results in rules deduction.