Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs

Ekhlas Sonu, Y. (Yingke) Chen, P. (Prashant) Doshi

Research output: Contribution to conferencePaperResearchpeer-review

Abstract

Interactive partially observable Markov decision processes (I-POMDP) provide a formal framework for planning for a self-interested agent in multiagent settings. An agent operating in a multiagent environment must deliberate about the actions that other agents may take and the effect these actions have on the environment and the rewards it receives. Traditional I-POMDPs model this dependence on the actions of other agents using joint action and model spaces. Therefore, the solution complexity grows exponentially with the number of agents thereby complicating scalability. In this paper, we model and extend anonymity and context-specific independence problem structures often present in agent populations for computational gain. We empirically demonstrate the efficiency from exploiting these problem structures by solving a new multiagent problem involving more than 1,000 agents.
Original languageEnglish
Publication statusPublished - 8 Apr 2015
Event25th International Conference on Automated Planning and Scheduling - Jerusalem, Israel
Duration: 7 Jun 201511 Jun 2015

Conference

Conference25th International Conference on Automated Planning and Scheduling
CountryIsrael
CityJerusalem
Period7/06/1511/06/15

Fingerprint

Planning
Scalability

Bibliographical note

Can archive post-print (ie final draft post-refereeing) or publisher's version/PDF Copyright IAAA. For full details see https://www.aaai.org/ocs/index.php/ICAPS/ICAPS15/about/submissions#copyrightNotice [Accessed 23/02/2018]

Cite this

Sonu, E., Chen, Y. Y., & Doshi, P. P. (2015). Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs. Paper presented at 25th International Conference on Automated Planning and Scheduling, Jerusalem, Israel.
Sonu, Ekhlas ; Chen, Y. (Yingke) ; Doshi, P. (Prashant). / Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs. Paper presented at 25th International Conference on Automated Planning and Scheduling, Jerusalem, Israel.
@conference{28325ffc3a9944779a4f37ed830c6688,
title = "Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs",
abstract = "Interactive partially observable Markov decision processes (I-POMDP) provide a formal framework for planning for a self-interested agent in multiagent settings. An agent operating in a multiagent environment must deliberate about the actions that other agents may take and the effect these actions have on the environment and the rewards it receives. Traditional I-POMDPs model this dependence on the actions of other agents using joint action and model spaces. Therefore, the solution complexity grows exponentially with the number of agents thereby complicating scalability. In this paper, we model and extend anonymity and context-specific independence problem structures often present in agent populations for computational gain. We empirically demonstrate the efficiency from exploiting these problem structures by solving a new multiagent problem involving more than 1,000 agents.",
author = "Ekhlas Sonu and Chen, {Y. (Yingke)} and Doshi, {P. (Prashant)}",
note = "Can archive post-print (ie final draft post-refereeing) or publisher's version/PDF Copyright IAAA. For full details see https://www.aaai.org/ocs/index.php/ICAPS/ICAPS15/about/submissions#copyrightNotice [Accessed 23/02/2018]; 25th International Conference on Automated Planning and Scheduling ; Conference date: 07-06-2015 Through 11-06-2015",
year = "2015",
month = "4",
day = "8",
language = "English",

}

Sonu, E, Chen, YY & Doshi, PP 2015, 'Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs' Paper presented at 25th International Conference on Automated Planning and Scheduling, Jerusalem, Israel, 7/06/15 - 11/06/15, .

Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs. / Sonu, Ekhlas; Chen, Y. (Yingke); Doshi, P. (Prashant).

2015. Paper presented at 25th International Conference on Automated Planning and Scheduling, Jerusalem, Israel.

Research output: Contribution to conferencePaperResearchpeer-review

TY - CONF

T1 - Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs

AU - Sonu, Ekhlas

AU - Chen, Y. (Yingke)

AU - Doshi, P. (Prashant)

N1 - Can archive post-print (ie final draft post-refereeing) or publisher's version/PDF Copyright IAAA. For full details see https://www.aaai.org/ocs/index.php/ICAPS/ICAPS15/about/submissions#copyrightNotice [Accessed 23/02/2018]

PY - 2015/4/8

Y1 - 2015/4/8

N2 - Interactive partially observable Markov decision processes (I-POMDP) provide a formal framework for planning for a self-interested agent in multiagent settings. An agent operating in a multiagent environment must deliberate about the actions that other agents may take and the effect these actions have on the environment and the rewards it receives. Traditional I-POMDPs model this dependence on the actions of other agents using joint action and model spaces. Therefore, the solution complexity grows exponentially with the number of agents thereby complicating scalability. In this paper, we model and extend anonymity and context-specific independence problem structures often present in agent populations for computational gain. We empirically demonstrate the efficiency from exploiting these problem structures by solving a new multiagent problem involving more than 1,000 agents.

AB - Interactive partially observable Markov decision processes (I-POMDP) provide a formal framework for planning for a self-interested agent in multiagent settings. An agent operating in a multiagent environment must deliberate about the actions that other agents may take and the effect these actions have on the environment and the rewards it receives. Traditional I-POMDPs model this dependence on the actions of other agents using joint action and model spaces. Therefore, the solution complexity grows exponentially with the number of agents thereby complicating scalability. In this paper, we model and extend anonymity and context-specific independence problem structures often present in agent populations for computational gain. We empirically demonstrate the efficiency from exploiting these problem structures by solving a new multiagent problem involving more than 1,000 agents.

M3 - Paper

ER -

Sonu E, Chen YY, Doshi PP. Individual Planning in Agent Populations: Exploiting Anonymity and Frame-Action Hypergraphs. 2015. Paper presented at 25th International Conference on Automated Planning and Scheduling, Jerusalem, Israel.