Nonstationary system control using Q-learning
Abstract eng: Q-learning is the most popular and effective version of Reinforcement Learning algorithms. In this paper we discuss the possibility of control of a nonstationary system by Q-learning. The non-stationary system is represented by simple inverted pendulum simulation model with variable pendulum length.
Contributors:
Publisher:
Institute of Thermomechanics AS CR, v.v.i., Brno
Conference Title:
Conference Title:
ENGINEERING MECHANICS 2005
Conference Venue:
Svratka (CZ)
Conference Dates:
2005-05-09 / 2005-05-12
Rights:
Text je chráněný podle autorského zákona č. 121/2000 Sb.
Record appears in:
Record created 2014-11-12, last modified 2014-11-18