Learning policies for abstract state spaces
Autor(en): | Timmer, S. Riedmiller, M. |
Stichwörter: | Cart-pole system; Learning process; Q-function; State space, Approximation theory; Dynamic programming; Functions, Learning systems | Erscheinungsdatum: | 2005 | Journal: | Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics | Volumen: | 4 | Startseite: | 3179 | Seitenende: | 3184 | Zusammenfassung: | Applying Q-Learning to multidimensional, real-valued state spaces is time-consuming in most cases. In this article, we deal with the assumption that a coarse partition of the state space is sufficient for learning good or even optimal policies. An algorithm is presented which constructs proper policies for abstract state spaces using an incremental procedure without approximating a Q-function. By combining an approach similar to dynamic programming and a search for policies, we can speed up the learning process. To provide empirical evidence, we use a cart-pole system. Experiments were conducted for a simulated environment as well as for a real plant. © 2005 IEEE. |
Beschreibung: | Conference of IEEE Systems, Man and Cybernetics Society, Proceedings - 2005 International Conference on Systems, Man and Cybernetics ; Conference Date: 10 October 2005 Through 12 October 2005; Conference Code:66062 |
ISSN: | 1062922X | Externe URL: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-27944452968&partnerID=40&md5=e8dad867fe1a18ce4c9bfef988c42210 |
Zur Langanzeige