Reinforcement learning for DEC-MDPs with changing action sets and partially ordered dependencies
Autor(en): | Gabel, T. Riedmiller, M. |
Stichwörter: | Autonomous agents; Communication; Computational complexity; Decentralized Markov Decision Processes; Decentralized MDPs; High quality; Interaction; Markov processes; Multi-agent reinforcement learning; Reduced complexity; Reinforcement learning, Action sets; State transitions, Multi agent systems | Erscheinungsdatum: | 2008 | Herausgeber: | International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS) | Journal: | Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS | Volumen: | 3 | Startseite: | 1305 | Seitenende: | 1308 | Zusammenfassung: | Decentralized Markov decision processes are frequently used to model cooperative multi-agent systems. In this paper, we identify a subclass of general DEC-MDPs that features regularities in the way agents interact with one another. This class is of high relevance for many real-world applications and features provably reduced complexity (NP-complete) compared to the general problem (NEXP-complete). Since optimally solving larger-sized NP-hard problems is intractable, we keep the learning as much decentralized as possible and use multi-agent reinforcement learning to improve the agents' behavior online. Further, we suggest a restricted message passing scheme that notifies other agents about forthcoming effects on their state transitions and that al- lows the agents to acquire approximate joint policies of high quality. Copyright © 2008, International Foundation for Autonomous Agents and Multiagent Systems (www.ifaamas.org). All rights reserved. |
Beschreibung: | Conference of 7th International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS 2008 ; Conference Date: 12 May 2008 Through 16 May 2008; Conference Code:105064 |
ISBN: | 9781605604701 | ISSN: | 15488403 | Externe URL: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-84899931898&partnerID=40&md5=28c4cd894dd7fbde5616fdea6ba7d4e9 |
Show full item record