An evolutionary algorithm for fuzzy controller synthesis and optimization

This paper illustrates an evolutionary algorithm for the automatic synthesis and optimization of fuzzy controllers for the weight associative rule processor (WARP) manufactured by SGS-Thomson Microelectronics. The philosophy of this work was to exploit the full potential of such a processor rather than use very simple membership functions and rule structures to reduce the number of degrees of freedom in the search. Details of the algorithm implementation and a discussion of experimental results obtained for the ball-and-beam problem are given.