Publication detail

THE INFLUENCE OF REINFORCEMENT FUNCTION DEFINITION ON PROPERTIES OF Q-LEARNING BASED CONTROLLER FOR ACTIVE MAGNETIC BEARING

BŘEZINA, T., KREJSA, J.

Original Title

THE INFLUENCE OF REINFORCEMENT FUNCTION DEFINITION ON PROPERTIES OF Q-LEARNING BASED CONTROLLER FOR ACTIVE MAGNETIC BEARING

Type

conference paper

Language

English

Original Abstract

Reinforcement learning (RL) based methods can be successfully used for control of non-stable systems such as active magnetic bearing (AMB). The paper shows the use of two phase Q-learning for such task, mainly with respect to the reinforcement function definition. Presented tests proved that Q-learning based controllers overpass referential PID controller.

Keywords

Control, Q-learning, Active Magnetic Bearing

Authors

BŘEZINA, T., KREJSA, J.

RIV year

2003

Released

18. 6. 2003

Publisher

Alexander Dubček University of Trenčín, Faculty of Mechatronics, Trenčín, SLOVAKIA

Location

Trenčín, SK

ISBN

80-88914-92-2

Book

Proceedings of 6th international symposium on Mechatronics

Edition number

1

Pages from

62

Pages to

65

Pages count

4

BibTex

@inproceedings{BUT7887,
  author="Tomáš {Březina} and Jiří {Krejsa}",
  title="THE INFLUENCE OF REINFORCEMENT FUNCTION DEFINITION ON PROPERTIES OF Q-LEARNING BASED CONTROLLER FOR ACTIVE MAGNETIC BEARING",
  booktitle="Proceedings of 6th international symposium on Mechatronics",
  year="2003",
  number="1",
  pages="4",
  publisher="Alexander Dubček University of Trenčín, Faculty of Mechatronics, Trenčín, SLOVAKIA",
  address="Trenčín, SK",
  isbn="80-88914-92-2"
}