Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Archives of Control Sciences

1 2008

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Optimization of fuzzy PID controllers using Q-learning algorithm

Boubertakh H., Labiod S., Tadjine M., Glorennec P.-Y.

Archives of Control Sciences

2008

Vol. 18, no. 4

415-435

In this article, we first chose the design settings of the fuzzy PID controllers (FPIDC) so that the FPIDCs mimic the classical PID controllers. The advantage of these controllers is the combination of the simplicity of the classical PID controllers and the interpretability of fuzzy controllers which makes the task of parameters tuning easier. Secondly, we present a method for optimizing the closed-loop system consisting of a FPIDC and an unknown plant using the Q-learning algorithm (QLA). Specifically, QLA minimizes a cost function which quantifies the performance of FPIDC. Without loss of generality the square error sum cost function is used. The QLA, which is a nonmodel-based method, iteratively search of the best parameters so that the output of the cost function is less then satisfaction threshold. Finally, a simulation example is used to prove the effectiveness of the proposed method.