Monte-Carlo expectation maximization for decentralized POMDPs
Monte-Carlo expectation maximization for decentralized POMDPs
We address two significant drawbacks of state-of-the-art solvers of decentralized POMDPs (DECPOMDPs): the reliance on complete knowledge of the model and limited scalability as the complexity of the domain grows. We extend a recently proposed approach for solving DEC-POMDPs via a reduction to the maximum likelihood problem, which in turn can be solved using EM. We introduce a model-free version of this approach that employs Monte-Carlo EM (MCEM). While a naïve implementation of MCEM is inadequate in multiagent settings, we introduce several improvements in sampling that produce high-quality results on a variety of DEC-POMDP benchmarks, including large problems with thousands of agents.
978-1-57735-633-2
397-403
Wu, Feng
b79f9800-2819-40c8-96e7-3ad85f866f5e
Zilberstein, S
f4822d34-5343-490c-a9ed-7be38f6c06d8
Jennings, N.R.
ab3d94cc-247c-4545-9d1e-65873d6cdb30
2013
Wu, Feng
b79f9800-2819-40c8-96e7-3ad85f866f5e
Zilberstein, S
f4822d34-5343-490c-a9ed-7be38f6c06d8
Jennings, N.R.
ab3d94cc-247c-4545-9d1e-65873d6cdb30
Wu, Feng, Zilberstein, S and Jennings, N.R.
(2013)
Monte-Carlo expectation maximization for decentralized POMDPs.
Proceedings of the 23rd International Joint Conference on AI (IJCAI), Beijing Shi, China.
03 - 09 Aug 2013.
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
We address two significant drawbacks of state-of-the-art solvers of decentralized POMDPs (DECPOMDPs): the reliance on complete knowledge of the model and limited scalability as the complexity of the domain grows. We extend a recently proposed approach for solving DEC-POMDPs via a reduction to the maximum likelihood problem, which in turn can be solved using EM. We introduce a model-free version of this approach that employs Monte-Carlo EM (MCEM). While a naïve implementation of MCEM is inadequate in multiagent settings, we introduce several improvements in sampling that produce high-quality results on a variety of DEC-POMDP benchmarks, including large problems with thousands of agents.
Text
ijcai2013.pdf
- Author's Original
More information
Published date: 2013
Venue - Dates:
Proceedings of the 23rd International Joint Conference on AI (IJCAI), Beijing Shi, China, 2013-08-03 - 2013-08-09
Organisations:
Agents, Interactions & Complexity
Identifiers
Local EPrints ID: 351021
URI: http://eprints.soton.ac.uk/id/eprint/351021
ISBN: 978-1-57735-633-2
PURE UUID: eda73ca3-f14c-4c55-bbb0-bd371901d1e2
Catalogue record
Date deposited: 12 Apr 2013 15:41
Last modified: 14 Mar 2024 13:34
Export record
Contributors
Author:
Feng Wu
Author:
S Zilberstein
Author:
N.R. Jennings
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics