Online collaborative multi-agent reinforcement learning by transfer of abstract trajectories
| Authors |
|
|---|---|
| Publication date | 2008 |
| Journal | BNAIC |
| Event | Twentieth Belgian-Netherlands Conference on Artificial Intelligence (BNAIC 2008), Enschede, the Netherlands |
| Volume | Issue number | 20 |
| Pages (from-to) | 241-248 |
| Organisations |
|
| Abstract | In this paper we propose a method for multi-agent reinforcement learning by automatic discovery of abstract trajectories. Local details are abstracted from successful trajectories and the resulting generalized, abstract trajectories are exchanged between agents. Each agent learns a policy for its own environment. By abstracting trajectories and sharing the result the agents benefit from each others learning. This reduces the overall learning time compared to individual learning. |
| Document type | Article |
| Note |
Proceedings title: BNAIC 2008: Belgian-Dutch Conference on Artificial Intelligence: proceedings of the twentieth Belgian-Dutch Conference on Artificial Intelligence: Enschede, October 30-31, 2008 Publisher: Universiteit Twente, Faculteit Elektrotechniek, Wiskunde en Informatica Place of publication: Enschede Editors: A. Nijholt, M. Pantic, M. Poel, H. Hondorp |
| Language | English |
| Published at | http://eprints.eemcs.utwente.nl/13354/ |
| Downloads | |
| Permalink to this page | |