Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward

Many cooperative multi-agent problems require agents to learn individual tasks while contributing to the collective success of the group. This is a challenging task for current state-of-the-art multi-agent reinforcement algorithms that are designed to either maximize the global reward of the team or...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sheikh, Hassam Ullah, Bölöni, Ladislau
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Sheikh, Hassam Ullah
Bölöni, Ladislau
description Many cooperative multi-agent problems require agents to learn individual tasks while contributing to the collective success of the group. This is a challenging task for current state-of-the-art multi-agent reinforcement algorithms that are designed to either maximize the global reward of the team or the individual local rewards. The problem is exacerbated when either of the rewards is sparse leading to unstable learning. To address this problem, we present Decomposed Multi-Agent Deep Deterministic Policy Gradient (DE-MADDPG): a novel cooperative multi-agent reinforcement learning framework that simultaneously learns to maximize the global and local rewards. We evaluate our solution on the challenging defensive escort team problem and show that our solution achieves a significantly better and more stable performance than the direct adaptation of the MADDPG algorithm.
doi_str_mv 10.48550/arxiv.2003.10598
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2003_10598</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2003_10598</sourcerecordid><originalsourceid>FETCH-LOGICAL-a678-208ff8bedf208b1f91c48161f5f2195359f9948a346e90ebdc53fe14d76ae8f63</originalsourceid><addsrcrecordid>eNotj7tOwzAUhr0woJYHYMIvkGDHcWqPVcSlUlARysAWncTntJYSB7lpC29PWpj-y_BJH2P3UqS50Vo8Qvz2pzQTQqVSaGtu2efbsZ98st5hmPgH-kBj7HC4rAohBh92fL74exzbHocDP_tpz8txaH1AxzfB-ZN3R-g5BMdrhGGmnCG6Jbsh6A94958LVj8_1eVrUm1fNuW6SqBYmSQThsi06GhurSQru9zIQpKmTFqttCVrcwMqL9AKbF2nFaHM3aoANFSoBXv4w17Vmq_oB4g_zUWxuSqqX_nPTHg</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward</title><source>arXiv.org</source><creator>Sheikh, Hassam Ullah ; Bölöni, Ladislau</creator><creatorcontrib>Sheikh, Hassam Ullah ; Bölöni, Ladislau</creatorcontrib><description>Many cooperative multi-agent problems require agents to learn individual tasks while contributing to the collective success of the group. This is a challenging task for current state-of-the-art multi-agent reinforcement algorithms that are designed to either maximize the global reward of the team or the individual local rewards. The problem is exacerbated when either of the rewards is sparse leading to unstable learning. To address this problem, we present Decomposed Multi-Agent Deep Deterministic Policy Gradient (DE-MADDPG): a novel cooperative multi-agent reinforcement learning framework that simultaneously learns to maximize the global and local rewards. We evaluate our solution on the challenging defensive escort team problem and show that our solution achieves a significantly better and more stable performance than the direct adaptation of the MADDPG algorithm.</description><identifier>DOI: 10.48550/arxiv.2003.10598</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning ; Computer Science - Multiagent Systems</subject><creationdate>2020-03</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2003.10598$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2003.10598$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Sheikh, Hassam Ullah</creatorcontrib><creatorcontrib>Bölöni, Ladislau</creatorcontrib><title>Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward</title><description>Many cooperative multi-agent problems require agents to learn individual tasks while contributing to the collective success of the group. This is a challenging task for current state-of-the-art multi-agent reinforcement algorithms that are designed to either maximize the global reward of the team or the individual local rewards. The problem is exacerbated when either of the rewards is sparse leading to unstable learning. To address this problem, we present Decomposed Multi-Agent Deep Deterministic Policy Gradient (DE-MADDPG): a novel cooperative multi-agent reinforcement learning framework that simultaneously learns to maximize the global and local rewards. We evaluate our solution on the challenging defensive escort team problem and show that our solution achieves a significantly better and more stable performance than the direct adaptation of the MADDPG algorithm.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><subject>Computer Science - Multiagent Systems</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj7tOwzAUhr0woJYHYMIvkGDHcWqPVcSlUlARysAWncTntJYSB7lpC29PWpj-y_BJH2P3UqS50Vo8Qvz2pzQTQqVSaGtu2efbsZ98st5hmPgH-kBj7HC4rAohBh92fL74exzbHocDP_tpz8txaH1AxzfB-ZN3R-g5BMdrhGGmnCG6Jbsh6A94958LVj8_1eVrUm1fNuW6SqBYmSQThsi06GhurSQru9zIQpKmTFqttCVrcwMqL9AKbF2nFaHM3aoANFSoBXv4w17Vmq_oB4g_zUWxuSqqX_nPTHg</recordid><startdate>20200323</startdate><enddate>20200323</enddate><creator>Sheikh, Hassam Ullah</creator><creator>Bölöni, Ladislau</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20200323</creationdate><title>Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward</title><author>Sheikh, Hassam Ullah ; Bölöni, Ladislau</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a678-208ff8bedf208b1f91c48161f5f2195359f9948a346e90ebdc53fe14d76ae8f63</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><topic>Computer Science - Multiagent Systems</topic><toplevel>online_resources</toplevel><creatorcontrib>Sheikh, Hassam Ullah</creatorcontrib><creatorcontrib>Bölöni, Ladislau</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Sheikh, Hassam Ullah</au><au>Bölöni, Ladislau</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward</atitle><date>2020-03-23</date><risdate>2020</risdate><abstract>Many cooperative multi-agent problems require agents to learn individual tasks while contributing to the collective success of the group. This is a challenging task for current state-of-the-art multi-agent reinforcement algorithms that are designed to either maximize the global reward of the team or the individual local rewards. The problem is exacerbated when either of the rewards is sparse leading to unstable learning. To address this problem, we present Decomposed Multi-Agent Deep Deterministic Policy Gradient (DE-MADDPG): a novel cooperative multi-agent reinforcement learning framework that simultaneously learns to maximize the global and local rewards. We evaluate our solution on the challenging defensive escort team problem and show that our solution achieves a significantly better and more stable performance than the direct adaptation of the MADDPG algorithm.</abstract><doi>10.48550/arxiv.2003.10598</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2003.10598
ispartof
issn
language eng
recordid cdi_arxiv_primary_2003_10598
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
Computer Science - Multiagent Systems
title Multi-Agent Reinforcement Learning for Problems with Combined Individual and Team Reward
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-17T23%3A22%3A05IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Multi-Agent%20Reinforcement%20Learning%20for%20Problems%20with%20Combined%20Individual%20and%20Team%20Reward&rft.au=Sheikh,%20Hassam%20Ullah&rft.date=2020-03-23&rft_id=info:doi/10.48550/arxiv.2003.10598&rft_dat=%3Carxiv_GOX%3E2003_10598%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true