Verdegay. Design of a control rules base based on genetic algorithms.pdf

DESIGN OF A CONTROL RULES BASE BASED ON GENETIC ALGORITHMS Francisco HERRERA Manuel LOZANO Jose Luis VERDEGAY Dept of Computer Science and Articial Intelligence University of Granada Granada Spain email herrera lozano verdegayrobinsonugres Abstract The purp ose is to present a learning pro cess from examples by means of genetic algorithms The learning pro cess is based on a generation metho d of fuzzy rules a simplify metho d which obtains the numb er of rules necessary to approximate a real system and a tuning metho d of control rules All parts are based on the use of dierent genetic algorithms whose chromosomes represent the adecuated information Keywords Fuzzy Control Rules Learning Genetic Algorithms Intro duction complex spaces oering a valid approach to prob lems requiring ecient and eective search Fuzzy Rule based mo dels are widthly used in au Our purp ose is to present a pro cess for learning tomation and engineering Fuzzy rules based sys fuzzy control rules based on the use of GAs under tems have b een shown to b e an imp ortant to ol for the following hyp otheses mo delling complex systems in which due to the complexity or the imprecision classical to ols are There is linguistic information from the human unsuccessful controllers exp erience But linguistic rules The exp erience of skilled op erators and the alone are usually not enough for designing a knowledge of control engineers are expressed quali successfully control system or might not b e tatively by a set of fuzzy control rules In fact one available of the features of the fuzzy logic controllers FLCs is that the IFTHEN rules are describ ed on the There is numerical information from sam ba