Multi-Agent deep reinforcement learning framework and method based on state classification and assignment

The invention belongs to the field of artificial intelligence, and particularly relates to a multi-agent deep reinforcement learning framework and method based on state classification and assignment. Comprising the steps that M1 is executed, generated environment virtual state samples are classified...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: XU JIN, BU JINFENG
Format: Patent
Sprache:chi ; 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 XU JIN
BU JINFENG
description The invention belongs to the field of artificial intelligence, and particularly relates to a multi-agent deep reinforcement learning framework and method based on state classification and assignment. Comprising the steps that M1 is executed, generated environment virtual state samples are classified, an assignment problem is solved, and a state classification model F and an assignment rule A of an agent are obtained; the method comprises the following steps: setting network structures of all agents in Multi-Agents, and learning modes, action selection strategies, updating modes and reward calculation methods of all agents in Multi-Agents; and executing M2, firstly initializing a plurality of homogeneous agents conforming to the setting of the above steps, then initializing the environment, and finally interacting the plurality of agents with the environment, learning action strategies and updating respective networks. According to the method, a complex environment state is decomposed into a plurality of simpl
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN115563527A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN115563527A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN115563527A3</originalsourceid><addsrcrecordid>eNqNzLEKwjAUheEuDqK-w_UBOtRSnUtRXHRyL9fkpAbTm5BEfH0N-ABOBz5-zrKyl5fLtu4nSCYNBIqwYnxUmAs5cBQrE5nIM94-PolF04z88JrunKDJC6XMGaQcp2SNVZztF0tYYJJyta4Whl3C5rerans63oZzjeBHpMAKgjwO16bpun3b7Q59-0_zAXDiQRM</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Multi-Agent deep reinforcement learning framework and method based on state classification and assignment</title><source>esp@cenet</source><creator>XU JIN ; BU JINFENG</creator><creatorcontrib>XU JIN ; BU JINFENG</creatorcontrib><description>The invention belongs to the field of artificial intelligence, and particularly relates to a multi-agent deep reinforcement learning framework and method based on state classification and assignment. Comprising the steps that M1 is executed, generated environment virtual state samples are classified, an assignment problem is solved, and a state classification model F and an assignment rule A of an agent are obtained; the method comprises the following steps: setting network structures of all agents in Multi-Agents, and learning modes, action selection strategies, updating modes and reward calculation methods of all agents in Multi-Agents; and executing M2, firstly initializing a plurality of homogeneous agents conforming to the setting of the above steps, then initializing the environment, and finally interacting the plurality of agents with the environment, learning action strategies and updating respective networks. According to the method, a complex environment state is decomposed into a plurality of simpl</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2023</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230103&amp;DB=EPODOC&amp;CC=CN&amp;NR=115563527A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,776,881,25543,76294</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230103&amp;DB=EPODOC&amp;CC=CN&amp;NR=115563527A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>XU JIN</creatorcontrib><creatorcontrib>BU JINFENG</creatorcontrib><title>Multi-Agent deep reinforcement learning framework and method based on state classification and assignment</title><description>The invention belongs to the field of artificial intelligence, and particularly relates to a multi-agent deep reinforcement learning framework and method based on state classification and assignment. Comprising the steps that M1 is executed, generated environment virtual state samples are classified, an assignment problem is solved, and a state classification model F and an assignment rule A of an agent are obtained; the method comprises the following steps: setting network structures of all agents in Multi-Agents, and learning modes, action selection strategies, updating modes and reward calculation methods of all agents in Multi-Agents; and executing M2, firstly initializing a plurality of homogeneous agents conforming to the setting of the above steps, then initializing the environment, and finally interacting the plurality of agents with the environment, learning action strategies and updating respective networks. According to the method, a complex environment state is decomposed into a plurality of simpl</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2023</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNzLEKwjAUheEuDqK-w_UBOtRSnUtRXHRyL9fkpAbTm5BEfH0N-ABOBz5-zrKyl5fLtu4nSCYNBIqwYnxUmAs5cBQrE5nIM94-PolF04z88JrunKDJC6XMGaQcp2SNVZztF0tYYJJyta4Whl3C5rerans63oZzjeBHpMAKgjwO16bpun3b7Q59-0_zAXDiQRM</recordid><startdate>20230103</startdate><enddate>20230103</enddate><creator>XU JIN</creator><creator>BU JINFENG</creator><scope>EVB</scope></search><sort><creationdate>20230103</creationdate><title>Multi-Agent deep reinforcement learning framework and method based on state classification and assignment</title><author>XU JIN ; BU JINFENG</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN115563527A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2023</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>XU JIN</creatorcontrib><creatorcontrib>BU JINFENG</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>XU JIN</au><au>BU JINFENG</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Multi-Agent deep reinforcement learning framework and method based on state classification and assignment</title><date>2023-01-03</date><risdate>2023</risdate><abstract>The invention belongs to the field of artificial intelligence, and particularly relates to a multi-agent deep reinforcement learning framework and method based on state classification and assignment. Comprising the steps that M1 is executed, generated environment virtual state samples are classified, an assignment problem is solved, and a state classification model F and an assignment rule A of an agent are obtained; the method comprises the following steps: setting network structures of all agents in Multi-Agents, and learning modes, action selection strategies, updating modes and reward calculation methods of all agents in Multi-Agents; and executing M2, firstly initializing a plurality of homogeneous agents conforming to the setting of the above steps, then initializing the environment, and finally interacting the plurality of agents with the environment, learning action strategies and updating respective networks. According to the method, a complex environment state is decomposed into a plurality of simpl</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN115563527A
source esp@cenet
subjects CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title Multi-Agent deep reinforcement learning framework and method based on state classification and assignment
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-23T02%3A31%3A56IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=XU%20JIN&rft.date=2023-01-03&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN115563527A%3C/epo_EVB%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