Using OGA in fuzzy based system modeling
Navid Seifipour, Mohammad Bagher Menhaj · 2002
High performance of fuzzy systems for modeling depends strongly on some parameters such as number of fuzzy partitions, their shapes and characteristics of membership functions. These parameters are usually chosen intuitively or more possibly after some trial-and-errors. This paper presents two techniques using a modified genetic algorithm and Marquardt BP based learning algorithm to improve fuzzy system models by systematically tuning the aforementioned parameters. To illustrate the effectiveness of the proposed technique, we employ them to model a synchronous generator. The simulation results are promising.