Design of a fuzzy logic controller for a plant of N-order based on genetic algorithms
Mohanad Alata, Mohammad Molhim, Khaled Al Masri · 2011
Design of fuzzy controllers has been always a job built on past experience and knowledge of fuzzy control and systems behavior. Unlike that fashion this research introduced a new methodology of designing fuzzy controllers using genetic algorithms. The design employs Sugeno type fuzzy controllers as the parameters can be manipulated using GA. Single and two inputs fuzzy controllers are used. This research presented a solution of first, second, and third order systems, using the absolute average error as a fitness function, the genetic algorithm manipulate all parameters of the fuzzy controller to find the optimum solution. The simulation model used Matlab GA toolbox for finding the optimal solution, the fitness function took a different shape than the usual form, the new shape is introduced using a short program that is capable of generating the whole system, then calculating its output, error and finally the average error which is used as a fitness value to finally design the fuzzy controller.