TY - GEN
T1 - Integrating design stages of fuzzy systems using genetic algorithms
AU - Lee, Michael A.
AU - Takagi, Hideyuki
PY - 1993
Y1 - 1993
N2 - This paper proposes an automatic fuzzy system design method that uses a Genetic Algorithm and integrates three design stages; our method determines membership functions, the number of fuzzy rules, and the rule-consequent parameters at the same time. Because these design stages may not be independent, it is important to consider them simultaneously to obtain optimal fuzzy systems. The method includes a genetic algorithm and a penalty strategy that favors systems with fewer rules. The proposed method is applied to the classic inverted pendulum control problem and has been shown to be practical through a comparison with another method.
AB - This paper proposes an automatic fuzzy system design method that uses a Genetic Algorithm and integrates three design stages; our method determines membership functions, the number of fuzzy rules, and the rule-consequent parameters at the same time. Because these design stages may not be independent, it is important to consider them simultaneously to obtain optimal fuzzy systems. The method includes a genetic algorithm and a penalty strategy that favors systems with fewer rules. The proposed method is applied to the classic inverted pendulum control problem and has been shown to be practical through a comparison with another method.
UR - http://www.scopus.com/inward/record.url?scp=0027224649&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0027224649&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:0027224649
SN - 0780306155
T3 - 1993 IEEE International Conference on Fuzzy Systems
SP - 612
EP - 617
BT - 1993 IEEE International Conference on Fuzzy Systems
PB - Publ by IEEE
T2 - Second IEEE International Conference on Fuzzy Systems
Y2 - 28 March 1993 through 1 April 1993
ER -