Standard

Bounding the probability of resource constraint violations in multi-agent MDPs. / De Nijs, Frits; Walraven, Erwin; De Weerdt, Mathijs M.; Spaan, Matthijs T.J.

Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017. American Association for Artificial Intelligence (AAAI), 2017. p. 3562-3568.

Research output: Scientific - peer-reviewConference contribution

Harvard

De Nijs, F, Walraven, E, De Weerdt, MM & Spaan, MTJ 2017, Bounding the probability of resource constraint violations in multi-agent MDPs. in Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017. American Association for Artificial Intelligence (AAAI), pp. 3562-3568, 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States, 4-10 February.

APA

De Nijs, F., Walraven, E., De Weerdt, M. M., & Spaan, M. T. J. (2017). Bounding the probability of resource constraint violations in multi-agent MDPs. In Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017. (pp. 3562-3568). American Association for Artificial Intelligence (AAAI).

Vancouver

De Nijs F, Walraven E, De Weerdt MM, Spaan MTJ. Bounding the probability of resource constraint violations in multi-agent MDPs. In Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017. American Association for Artificial Intelligence (AAAI). 2017. p. 3562-3568.

Author

De Nijs, Frits; Walraven, Erwin; De Weerdt, Mathijs M.; Spaan, Matthijs T.J. / Bounding the probability of resource constraint violations in multi-agent MDPs.

Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017. American Association for Artificial Intelligence (AAAI), 2017. p. 3562-3568.

Research output: Scientific - peer-reviewConference contribution

BibTeX

@inbook{5d6c967ee4b34a6580ec53d191e3da03,
title = "Bounding the probability of resource constraint violations in multi-agent MDPs",
author = "{De Nijs}, Frits and Erwin Walraven and {De Weerdt}, {Mathijs M.} and Spaan, {Matthijs T.J.}",
year = "2017",
pages = "3562--3568",
booktitle = "Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017",
publisher = "American Association for Artificial Intelligence (AAAI)",
address = "United States",

}

RIS

TY - CHAP

T1 - Bounding the probability of resource constraint violations in multi-agent MDPs

AU - De Nijs,Frits

AU - Walraven,Erwin

AU - De Weerdt,Mathijs M.

AU - Spaan,Matthijs T.J.

PY - 2017

Y1 - 2017

N2 - Multi-agent planning problems with constraints on global resource consumption occur in several domains. Existing algorithms for solving Multi-agent Markov Decision Processes can compute policies that meet a resource constraint in expectation, but these policies provide no guarantees on the probability that a resource constraint violation will occur. We derive a method to bound constraint violation probabilities using Hoeffding's inequality. This method is applied to two existing approaches for computing policies satisfying constraints: the Constrained MDP framework and a Column Generation approach. We also introduce an algorithm to adaptively relax the bound up to a given maximum violation tolerance. Experiments on a hard toy problem show that the resulting policies outperform static optimal resource allocations to an arbitrary level. By testing the algorithms on more realistic planning domains from the literature, we demonstrate that the adaptive bound is able to efficiently trade off violation probability with expected value, outperforming state-of-the-art planners.

AB - Multi-agent planning problems with constraints on global resource consumption occur in several domains. Existing algorithms for solving Multi-agent Markov Decision Processes can compute policies that meet a resource constraint in expectation, but these policies provide no guarantees on the probability that a resource constraint violation will occur. We derive a method to bound constraint violation probabilities using Hoeffding's inequality. This method is applied to two existing approaches for computing policies satisfying constraints: the Constrained MDP framework and a Column Generation approach. We also introduce an algorithm to adaptively relax the bound up to a given maximum violation tolerance. Experiments on a hard toy problem show that the resulting policies outperform static optimal resource allocations to an arbitrary level. By testing the algorithms on more realistic planning domains from the literature, we demonstrate that the adaptive bound is able to efficiently trade off violation probability with expected value, outperforming state-of-the-art planners.

UR - http://www.scopus.com/inward/record.url?scp=85030478873&partnerID=8YFLogxK

UR - https://aaai.org/ocs/index.php/AAAI/AAAI17/paper/view/15009

M3 - Conference contribution

SP - 3562

EP - 3568

BT - Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017

PB - American Association for Artificial Intelligence (AAAI)

ER -

ID: 29530600