Decentralized partially observable Markov decision processes (DEC-POMDPs) form a general framework for planning for groups of cooperating agents that inhabit a stochastic and partially observable environment. Unfortunately, computing optimal plans in a DEC-POMDP has been shown to be intractable (NEXP-complete), and approximate algorithms for specific subclasses have been proposed. Many of these algorithms rely on an (approximate) solution of the centralized planning problem (i.e., treating the whole team as a single agent). We take a more decentralized approach, in which each agent only reasons over its own local state and some uncontrollable state features, which are shared by all team members. In contrast to other approaches, we model communication as an integral part of the agent's reasoning, in which the meaning of a message is directly encoded in the policy of the communicating agent. We explore iterative methods for approximately solving such models, and we conclude with some encouraging preliminary experimental results.
%0 Conference Paper
%1 conf/aamas06/Spaan
%A Spaan, Matthijs T. J.
%A Gordon, Geoffrey J.
%A Vlassis, Nikos
%B AAMAS '06: Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
%C New York, NY, USA
%D 2006
%I ACM Press
%K mas planning pomdp toberead
%P 249--256
%R http://doi.acm.org/10.1145/1160633.1160678
%T Decentralized planning under uncertainty for teams of communicating agents
%U http://portal.acm.org/citation.cfm?id=1160633.1160678&coll=ACM&dl=ACM&type=series&idx=1160633&part=Proceedings&WantType=Proceedings&title=International%20Conference%20on%20Autonomous%20Agents&CFID=25754815&CFTOKEN=92537800
%X Decentralized partially observable Markov decision processes (DEC-POMDPs) form a general framework for planning for groups of cooperating agents that inhabit a stochastic and partially observable environment. Unfortunately, computing optimal plans in a DEC-POMDP has been shown to be intractable (NEXP-complete), and approximate algorithms for specific subclasses have been proposed. Many of these algorithms rely on an (approximate) solution of the centralized planning problem (i.e., treating the whole team as a single agent). We take a more decentralized approach, in which each agent only reasons over its own local state and some uncontrollable state features, which are shared by all team members. In contrast to other approaches, we model communication as an integral part of the agent's reasoning, in which the meaning of a message is directly encoded in the policy of the communicating agent. We explore iterative methods for approximately solving such models, and we conclude with some encouraging preliminary experimental results.
%@ 1-59593-303-4
@inproceedings{conf/aamas06/Spaan,
abstract = {Decentralized partially observable Markov decision processes (DEC-POMDPs) form a general framework for planning for groups of cooperating agents that inhabit a stochastic and partially observable environment. Unfortunately, computing optimal plans in a DEC-POMDP has been shown to be intractable (NEXP-complete), and approximate algorithms for specific subclasses have been proposed. Many of these algorithms rely on an (approximate) solution of the centralized planning problem (i.e., treating the whole team as a single agent). We take a more decentralized approach, in which each agent only reasons over its own local state and some uncontrollable state features, which are shared by all team members. In contrast to other approaches, we model communication as an integral part of the agent's reasoning, in which the meaning of a message is directly encoded in the policy of the communicating agent. We explore iterative methods for approximately solving such models, and we conclude with some encouraging preliminary experimental results.},
added-at = {2007-08-03T13:58:43.000+0200},
address = {New York, NY, USA},
author = {Spaan, Matthijs T. J. and Gordon, Geoffrey J. and Vlassis, Nikos},
biburl = {https://www.bibsonomy.org/bibtex/26a3effba01123a33041b341edafa2a19/mpfingst},
booktitle = {AAMAS '06: Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems},
doi = {http://doi.acm.org/10.1145/1160633.1160678},
interhash = {fd3add1903bea057f217fec66866face},
intrahash = {6a3effba01123a33041b341edafa2a19},
isbn = {1-59593-303-4},
keywords = {mas planning pomdp toberead},
location = {Hakodate, Japan},
pages = {249--256},
publisher = {ACM Press},
timestamp = {2007-08-03T13:58:43.000+0200},
title = {Decentralized planning under uncertainty for teams of communicating agents},
url = {http://portal.acm.org/citation.cfm?id=1160633.1160678&coll=ACM&dl=ACM&type=series&idx=1160633&part=Proceedings&WantType=Proceedings&title=International%20Conference%20on%20Autonomous%20Agents&CFID=25754815&CFTOKEN=92537800},
year = 2006
}