A Value Equivalence Approach for Solving Interactive Dynamic Influence Diagrams

Hdl Handle:
http://hdl.handle.net/10149/600590
Title:
A Value Equivalence Approach for Solving Interactive Dynamic Influence Diagrams
Authors:
Conroy, R. (Ross); Zeng, Y. (Yifeng); Cavazza, M. (Marc); Tang, J. (Jing); Pan, Y. (Yinghui)
Affiliation:
Teesside University, Digital Futures Institute
Citation:
Ross Conroy, Yifeng Zeng, Marc Cavazza, Jing Tang, Yinghui Pan (2016) 'A Value Equivalence Approach for Solving Interactive Dynamic Influence Diagrams' Proceedings of the 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS2016), 9-13 May 2016, Grand Copthorne Waterfront Hotel, Singapore
Publisher:
ACM
Journal:
Proceedings of the 15th International Conference on Autonomous Agents and Multiagent Systems
Conference:
the 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS2016), 9-13 May 2016, Grand Copthorne Waterfront Hotel, Singapore
Issue Date:
9-May-2016
URI:
http://hdl.handle.net/10149/600590
Additional Links:
http://sis.smu.edu.sg/aamas2016?itemid=671
Abstract:
Interactive dynamic influence diagrams (I-DIDs) are recognized graphical models for sequential multiagent decision making under uncertainty. They represent the problem of how a subject agent acts in a common setting shared with other agents who may act in sophisticated ways. The difficulty in solving I-DIDs is mainly due to an exponentially growing space of candidate models ascribed to other agents over time. in order to minimize the model space, the previous I-DID techniques prune behaviorally equivalent models. In this paper, we challenge the minimal set of models and propose a value equivalence approach to further compress the model space. The new method reduces the space by additionally pruning behaviorally distinct models that result in the same expected value of the subject agent’s optimal policy. To achieve this, we propose to learn the value from available data particularly in practical applications of real-time strategy games. We demonstrate the performance of the new technique in two problem domains.
Type:
Meetings and Proceedings
Language:
en_US
Rights:
ACM allows authors' version of their own ACM-copyrighted work on their personal server or on severs belonging to their employers. For full details see http://www.acm.org/publications/policies/RightsResponsibilities [Accessed 04/03/2016]

Full metadata record

DC FieldValue Language
dc.contributor.authorConroy, R. (Ross)en
dc.contributor.authorZeng, Y. (Yifeng)en
dc.contributor.authorCavazza, M. (Marc)en
dc.contributor.authorTang, J. (Jing)en
dc.contributor.authorPan, Y. (Yinghui)en
dc.date.accessioned2016-03-04T13:38:08Zen
dc.date.available2016-03-04T13:38:08Zen
dc.date.issued2016-05-09en
dc.identifier.urihttp://hdl.handle.net/10149/600590en
dc.description.abstractInteractive dynamic influence diagrams (I-DIDs) are recognized graphical models for sequential multiagent decision making under uncertainty. They represent the problem of how a subject agent acts in a common setting shared with other agents who may act in sophisticated ways. The difficulty in solving I-DIDs is mainly due to an exponentially growing space of candidate models ascribed to other agents over time. in order to minimize the model space, the previous I-DID techniques prune behaviorally equivalent models. In this paper, we challenge the minimal set of models and propose a value equivalence approach to further compress the model space. The new method reduces the space by additionally pruning behaviorally distinct models that result in the same expected value of the subject agent’s optimal policy. To achieve this, we propose to learn the value from available data particularly in practical applications of real-time strategy games. We demonstrate the performance of the new technique in two problem domains.en
dc.language.isoen_USen
dc.publisherACMen
dc.relation.urlhttp://sis.smu.edu.sg/aamas2016?itemid=671en
dc.rightsACM allows authors' version of their own ACM-copyrighted work on their personal server or on severs belonging to their employers. For full details see http://www.acm.org/publications/policies/RightsResponsibilities [Accessed 04/03/2016]en
dc.titleA Value Equivalence Approach for Solving Interactive Dynamic Influence Diagramsen_US
dc.typeMeetings and Proceedingsen
dc.contributor.departmentTeesside University, Digital Futures Instituteen
dc.identifier.journalProceedings of the 15th International Conference on Autonomous Agents and Multiagent Systemsen
dc.identifier.conferencethe 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS2016), 9-13 May 2016, Grand Copthorne Waterfront Hotel, Singaporeen
or.citation.harvardRoss Conroy, Yifeng Zeng, Marc Cavazza, Jing Tang, Yinghui Pan (2016) 'A Value Equivalence Approach for Solving Interactive Dynamic Influence Diagrams' Proceedings of the 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS2016), 9-13 May 2016, Grand Copthorne Waterfront Hotel, Singaporeen
All Items in TeesRep are protected by copyright, with all rights reserved, unless otherwise indicated.