Reinforcement learning for MDPs with constraints

Autor(en): Geibel, Peter
Herausgeber: Furnkranz, J
Scheffer, T
Spiliopoulou, M
Stichwörter: Computer Science; Computer Science, Artificial Intelligence
Erscheinungsdatum: 2006
Herausgeber: SPRINGER-VERLAG BERLIN
Journal: MACHINE LEARNING: ECML 2006, PROCEEDINGS
LECTURE NOTES IN COMPUTER SCIENCE
Volumen: 4212
Startseite: 646
Seitenende: 653
Zusammenfassung: 
In this article, I will consider Markov Decision Processes with two criteria, each defined as the expected value of an infinite horizon cumulative return. The second criterion is either itself subject to an inequality constraint, or there is maximum allowable probability that the single returns violate the constraint. I describe and discuss three new reinforcement learning approaches for solving such control problems.
Beschreibung: 
17th European Conference on Machine Learning (ECML 2006), Berlin, GERMANY, SEP 18-22, 2006
ISBN: 9783540453758
ISSN: 03029743

Show full item record

Page view(s)

2
Last Week
0
Last month
2
checked on Feb 24, 2024

Google ScholarTM

Check

Altmetric