Extended Markov Games to learn multiple tasks in multi-agent reinforcement learning
File(s)
Published version
Author(s)
Leon, BG
Belardinelli, F
Type
Conference Paper
Abstract
The combination of Formal Methods with Reinforcement Learning (RL) has recently attracted interest as a way for single-agent RL to learn multiple-task specifications. In this paper we extend this convergence to multi-agent settings and formally define Extended Markov Games as a general mathematical model that allows multiple RL agents to concurrently learn various non-Markovian specifications. To introduce this new model we provide formal definitions and proofs as well as empirical tests of RL algorithms running on this framework. Specifically, we use our model to train two different logic-based multi-agent RL algorithms to solve diverse settings of non-Markovian co-safe LTL specifications.
Date Issued
2020
Date Acceptance
2020-08-29
Citation
Frontiers in Artificial Intelligence and Applications, 2020, 325, pp.139-146
ISBN
9781643681009
ISSN
0922-6389
Publisher
IOS Press
Start Page
139
End Page
146
Journal / Book Title
Frontiers in Artificial Intelligence and Applications
Volume
325
Copyright Statement
© 2020 The authors and IOS Press.
This article is published online with Open Access by IOS Press and distributed under the terms
of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0).
This article is published online with Open Access by IOS Press and distributed under the terms
of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0).
License URL
Identifier
https://ebooks.iospress.nl/publication/54881
Source
24th European Conference on Artificial Intelligence, 29 August–8 September 2020, Santiago de Compostela, Spain – Including 10th Conference on Prestigious Applications of Artificial Intelligence (PAIS 2020)
Publication Status
Published
Start Date
2020-08-29
Finish Date
2020-09-08
Coverage Spatial
Santiago de Compostela, Spain
Date Publish Online
2020