A Fully Data-Driven Approach for Realistic Traffic Signal Control Using Offline Reinforcement Learning

The optimization of traffic signal control (TSC) is critical for an efficient transportation system. In recent years, reinforcement learning (RL) techniques have emerged as a popular approach for TSC and show promising results for highly adaptive control. However, existing RL-based methods suffer fr...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Data science for Transportation Ročník 7; číslo 3; s. 25
Hlavní autoři:	Li, Jianxiong, Lin, Shichao, Shi, Tianyu, Tian, Chujie, Mei, Yu, Song, Jian, Zhan, Xianyuan, Li, Ruimin
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	Singapore Springer Nature Singapore 01.12.2025 Springer Nature B.V
Témata:	Adaptive control Computational Intelligence Data Mining and Knowledge Discovery Datasets Engineering Flow theory Machine learning Neural networks Optimization Regularization methods Simulation Simulators Traffic control Traffic flow Traffic information Traffic intersections Traffic signals Transportation systems Transportation Technology and Traffic Engineering Real-world traffic data Shockwave theory Offline reinforcement learning Traffic signal control
ISSN:	2948-135X, 2948-1368
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Abstract	The optimization of traffic signal control (TSC) is critical for an efficient transportation system. In recent years, reinforcement learning (RL) techniques have emerged as a popular approach for TSC and show promising results for highly adaptive control. However, existing RL-based methods suffer from serious real-world transferability issues and hardly have any successful deployments. The reasons for such failures are mostly due to the reliance on over-idealized traffic simulators for policy optimization, as well as using unrealistic fine-grained state observations and reward signals that are not directly obtainable from real-world sensors. In this paper, we propose a fully data-driven and simulator-free framework for realistic Traffic Signal Control. Specifically, we combine well-established traffic flow theory with machine learning to construct a reward inference model to infer the reward signals from coarse-grained traffic data. With the inferred rewards, we further propose a sample-efficient offline RL method to enable direct signal control policy learning from historical offline data sets of real-world intersections. To evaluate our approach, we collect historical traffic data from a real-world intersection, and develop a highly customized simulation environment that strictly follows real data characteristics. We demonstrate through extensive experiments that our approach achieves superior performance over conventional and offline RL baselines, and also enjoys much better real-world applicability.
AbstractList	The optimization of traffic signal control (TSC) is critical for an efficient transportation system. In recent years, reinforcement learning (RL) techniques have emerged as a popular approach for TSC and show promising results for highly adaptive control. However, existing RL-based methods suffer from serious real-world transferability issues and hardly have any successful deployments. The reasons for such failures are mostly due to the reliance on over-idealized traffic simulators for policy optimization, as well as using unrealistic fine-grained state observations and reward signals that are not directly obtainable from real-world sensors. In this paper, we propose a fully data-driven and simulator-free framework for realistic Traffic Signal Control. Specifically, we combine well-established traffic flow theory with machine learning to construct a reward inference model to infer the reward signals from coarse-grained traffic data. With the inferred rewards, we further propose a sample-efficient offline RL method to enable direct signal control policy learning from historical offline data sets of real-world intersections. To evaluate our approach, we collect historical traffic data from a real-world intersection, and develop a highly customized simulation environment that strictly follows real data characteristics. We demonstrate through extensive experiments that our approach achieves superior performance over conventional and offline RL baselines, and also enjoys much better real-world applicability.
ArticleNumber	25
Author	Zhan, Xianyuan Shi, Tianyu Mei, Yu Li, Ruimin Lin, Shichao Li, Jianxiong Tian, Chujie Song, Jian
Author_xml	– sequence: 1 givenname: Jianxiong surname: Li fullname: Li, Jianxiong organization: Tsinghua University – sequence: 2 givenname: Shichao surname: Lin fullname: Lin, Shichao organization: Tsinghua University – sequence: 3 givenname: Tianyu surname: Shi fullname: Shi, Tianyu organization: University of Toronto – sequence: 4 givenname: Chujie surname: Tian fullname: Tian, Chujie organization: Baidu Inc – sequence: 5 givenname: Yu surname: Mei fullname: Mei, Yu organization: Baidu Inc – sequence: 6 givenname: Jian surname: Song fullname: Song, Jian organization: Baidu Inc – sequence: 7 givenname: Xianyuan surname: Zhan fullname: Zhan, Xianyuan email: zhanxianyuan@air.tsinghua.edu.cn organization: Tsinghua University – sequence: 8 givenname: Ruimin surname: Li fullname: Li, Ruimin email: lrmin@tsinghua.edu.cn organization: Tsinghua University
BookMark	eNp9UE1LAzEUDFLBWvsHPAU8R_PV3eyxtFaFQkFb8BayaVJT0mxNtkL7642u6M3TPN6bGd7MJeiFJhgArgm-JRiXd4lTTgnCdIQwJqxCpzPQpxUXiLBC9H7n0esFGKa0xRjTCpO87QM7hrOD90c4Va1C0-g-TIDj_T42Sr9B20T4bJR3qXUaLqOyNuOL2wTl4aQJbWw8XCUXNnBhrXfBZLoLWabNzoQWzo2KIZ-vwLlVPpnhDw7Aana_nDyi-eLhaTKeI01EeUJUqJLjmjGluRpVFAuLTW3qtWBrzK0WnK9LYrlRtGCmUlxYSktLtaWirouaDcBN55sDvB9MauW2OcT8bZKMFoLl4LTKLNqxdGxSisbKfXQ7FY-SYPlVqewqlblS-V2pPGUR60Qpk8PGxD_rf1SfzgZ8Fg
Cites_doi	10.1145/3447556.3447565 10.1145/3357384.3358079 10.1007/978-1-4471-5113-5_3 10.1016/0191-2615(93)90039-D 10.1016/j.trb.2015.02.015 10.1016/j.trc.2019.08.014 10.1109/ITSC.2018.8569938 10.1016/j.trc.2020.102660 10.1098/rspa.1955.0089 10.1016/j.trc.2022.103728 10.1109/JAS.2016.7508798 10.1108/9780585475301 10.1007/s42421-023-00074-x 10.1109/TITS.2021.3070835 10.1007/s42421-024-00103-3 10.1145/3357384.3357900 10.1609/aaai.v34i04.5744 10.1145/3308558.3314139 10.1609/aaai.v39i20.35403 10.1016/j.trc.2009.04.005 10.1609/aaai.v36i4.20393 10.1016/0191-2615(91)90013-9 10.1109/ITSC.2019.8917451 10.1145/3219819.3220096 10.1007/s42421-020-00021-0 10.1145/3357384.3357902 10.1109/TITS.2021.3054625 10.1109/T-VT.1980.23833 10.1609/aaai.v34i01.5467 10.1109/TITS.2023.3263847 10.1109/TITS.2019.2901791 10.1016/j.trc.2021.103321 10.1109/TVT.2018.2890726 10.1287/opre.4.1.42 10.1061/(ASCE)0733-947X(2003)129:3(278) 10.1007/s42421-020-00029-6
ContentType	Journal Article
Copyright	The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.
Copyright_xml	– notice: The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. – notice: The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.
DBID	AAYXX CITATION JQ2
DOI	10.1007/s42421-025-00139-z
DatabaseName	CrossRef ProQuest Computer Science Collection
DatabaseTitle	CrossRef ProQuest Computer Science Collection
DatabaseTitleList	ProQuest Computer Science Collection
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering
EISSN	2948-1368
ExternalDocumentID	10_1007_s42421_025_00139_z
GrantInformation_xml	– fundername: National Key Research and Development Program of China under Grant grantid: 2022YFB2502904 – fundername: Baidu funderid: https://doi.org/10.13039/100018903
GroupedDBID	0R~ 406 AACDK AAJBT AASML AATNV ABAKF ABBRH ABDBE ABFSG ABJNI ABRTQ ACAOD ACPIV ACSTC ACZOJ AEFQL AEMSY AESKC AEZWR AFBBN AFDZB AFHIU AFOHR AGQEE AHWEU AIGIU AIXLP ALMA_UNASSIGNED_HOLDINGS AMXSW ATHPR DPUIP EBLON EBS FIGPU JZLTJ NPVJJ PT4 ROL RSV SJYHP SNE SOJ AAUYE AAYXX ABJCF AEUYN AFFHD AFKRA ARAPS BENPR BGLVJ BGNMA CCPQU CITATION HCIFZ K7- M4Y M7S NU0 PHGZM PHGZT PQGLB PTHSS JQ2
ID	FETCH-LOGICAL-c187z-28a740b33ac4a59208f0ebebd83d04fc844d71f4ea263e9a48f227f2cf28bb6b3
IEDL.DBID	RSV
ISSN	2948-135X
IngestDate	Wed Nov 05 10:19:52 EST 2025 Sat Nov 29 07:01:42 EST 2025 Tue Nov 04 01:14:17 EST 2025
IsPeerReviewed	true
IsScholarly	true
Issue	3
Keywords	Real-world traffic data Shockwave theory Offline reinforcement learning Traffic signal control
Language	English
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c187z-28a740b33ac4a59208f0ebebd83d04fc844d71f4ea263e9a48f227f2cf28bb6b3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
PQID	3268302929
PQPubID	7434373
ParticipantIDs	proquest_journals_3268302929 crossref_primary_10_1007_s42421_025_00139_z springer_journals_10_1007_s42421_025_00139_z
PublicationCentury	2000
PublicationDate	2025-12-01
PublicationDateYYYYMMDD	2025-12-01
PublicationDate_xml	– month: 12 year: 2025 text: 2025-12-01 day: 01
PublicationDecade	2020
PublicationPlace	Singapore
PublicationPlace_xml	– name: Singapore – name: London
PublicationTitle	Data science for Transportation
PublicationTitleAbbrev	Data Sci. Transp
PublicationYear	2025
Publisher	Springer Nature Singapore Springer Nature B.V
Publisher_xml	– name: Springer Nature Singapore – name: Springer Nature B.V
References	CM Bishop (139_CR5) 2006 139_CR44 139_CR42 139_CR40 JO Berger (139_CR4) 2013 PI Richards (139_CR46) 1956; 4 139_CR47 139_CR48 139_CR45 H Xu (139_CR59) 2022; 35 A Kumar (139_CR26) 2020; 33 L Li (139_CR29) 2016; 3 AG Sims (139_CR49) 1980; 29 BR Kiran (139_CR24) 2021; 23 Z Mo (139_CR39) 2022; 141 139_CR54 139_CR11 139_CR55 C Cai (139_CR6) 2009; 17 139_CR53 139_CR50 139_CR9 139_CR51 139_CR7 MJ Lighthill (139_CR31) 1955; 229 139_CR18 139_CR19 139_CR16 S Fujimoto (139_CR12) 2021; 34 139_CR14 139_CR58 X Zhan (139_CR64) 2020; 117 139_CR15 139_CR56 139_CR13 139_CR57 A Oroojlooy (139_CR43) 2020; 33 139_CR21 139_CR65 139_CR22 139_CR66 139_CR63 139_CR20 NH Gartner (139_CR17) 1991; 25 139_CR60 139_CR27 139_CR28 H Wei (139_CR52) 2021; 22 139_CR25 H Yan (139_CR61) 2019; 107 139_CR69 139_CR67 139_CR68 T Chu (139_CR8) 2019; 21 139_CR3 139_CR32 139_CR76 139_CR33 139_CR74 139_CR75 B Abdulhai (139_CR1) 2003; 129 139_CR72 139_CR73 139_CR70 139_CR71 J Yoon (139_CR62) 2021; 130 139_CR38 139_CR36 139_CR37 139_CR34 139_CR35 GF Newell (139_CR41) 1993; 27 CF Daganzo (139_CR10) 1997 W-L Jin (139_CR23) 2015; 77 G An (139_CR2) 2021; 34 X Liang (139_CR30) 2019; 68
References_xml	– ident: 139_CR60 – ident: 139_CR58 – volume: 22 start-page: 12 issue: 2 year: 2021 ident: 139_CR52 publication-title: ACM SIGKDD Explor Newsl doi: 10.1145/3447556.3447565 – ident: 139_CR16 – ident: 139_CR57 doi: 10.1145/3357384.3358079 – ident: 139_CR9 doi: 10.1007/978-1-4471-5113-5_3 – volume: 27 start-page: 289 issue: 4 year: 1993 ident: 139_CR41 publication-title: Transp Res B Methodol doi: 10.1016/0191-2615(93)90039-D – volume: 34 start-page: 7436 year: 2021 ident: 139_CR2 publication-title: Adv Neural Inf Process Syst – ident: 139_CR21 – ident: 139_CR44 – volume: 77 start-page: 1 year: 2015 ident: 139_CR23 publication-title: Transp Res B Methodol doi: 10.1016/j.trb.2015.02.015 – ident: 139_CR25 – ident: 139_CR50 – ident: 139_CR48 – volume: 107 start-page: 266 year: 2019 ident: 139_CR61 publication-title: Transp Res C Emerg Technol doi: 10.1016/j.trc.2019.08.014 – ident: 139_CR35 doi: 10.1109/ITSC.2018.8569938 – ident: 139_CR40 – volume: 117 year: 2020 ident: 139_CR64 publication-title: Transp Res C Emerg Technol doi: 10.1016/j.trc.2020.102660 – volume: 229 start-page: 317 issue: 1178 year: 1955 ident: 139_CR31 publication-title: Proc R Soc Lond A doi: 10.1098/rspa.1955.0089 – volume: 141 year: 2022 ident: 139_CR39 publication-title: Transp Res C Emerg Technol doi: 10.1016/j.trc.2022.103728 – ident: 139_CR3 – ident: 139_CR34 – volume: 3 start-page: 247 issue: 3 year: 2016 ident: 139_CR29 publication-title: IEEE/CAA J Autom Sin doi: 10.1109/JAS.2016.7508798 – ident: 139_CR13 – volume-title: Fundamentals of transportation and traffic operations year: 1997 ident: 139_CR10 doi: 10.1108/9780585475301 – ident: 139_CR71 doi: 10.1007/s42421-023-00074-x – ident: 139_CR55 – ident: 139_CR38 – ident: 139_CR11 doi: 10.1109/TITS.2021.3070835 – ident: 139_CR70 doi: 10.1007/s42421-024-00103-3 – ident: 139_CR67 doi: 10.1145/3357384.3357900 – ident: 139_CR7 doi: 10.1609/aaai.v34i04.5744 – volume: 34 start-page: 20132 year: 2021 ident: 139_CR12 publication-title: Adv Neural Inf Process Syst – volume: 35 start-page: 4085 year: 2022 ident: 139_CR59 publication-title: Adv Neural Inf Process Syst – ident: 139_CR65 doi: 10.1145/3308558.3314139 – ident: 139_CR51 – ident: 139_CR45 – ident: 139_CR74 doi: 10.1609/aaai.v39i20.35403 – volume: 17 start-page: 456 issue: 5 year: 2009 ident: 139_CR6 publication-title: Transp Res C Emerg Technol doi: 10.1016/j.trc.2009.04.005 – ident: 139_CR28 – ident: 139_CR76 – volume-title: Statistical decision theory and Bayesian analysis year: 2013 ident: 139_CR4 – ident: 139_CR20 – volume: 33 start-page: 1179 year: 2020 ident: 139_CR26 publication-title: Adv Neural Inf Process Syst – ident: 139_CR69 doi: 10.1609/aaai.v36i4.20393 – volume: 25 start-page: 55 issue: 1 year: 1991 ident: 139_CR17 publication-title: Transp Res B Methodol doi: 10.1016/0191-2615(91)90013-9 – ident: 139_CR47 doi: 10.1109/ITSC.2019.8917451 – ident: 139_CR53 doi: 10.1145/3219819.3220096 – ident: 139_CR56 – ident: 139_CR33 – ident: 139_CR66 – ident: 139_CR14 – ident: 139_CR73 doi: 10.1007/s42421-020-00021-0 – ident: 139_CR54 doi: 10.1145/3357384.3357902 – volume: 23 start-page: 4909 issue: 6 year: 2021 ident: 139_CR24 publication-title: IEEE Trans Intell Transp Syst doi: 10.1109/TITS.2021.3054625 – volume: 29 start-page: 130 issue: 2 year: 1980 ident: 139_CR49 publication-title: IEEE Trans Veh Technol doi: 10.1109/T-VT.1980.23833 – ident: 139_CR75 – ident: 139_CR18 – ident: 139_CR27 – ident: 139_CR42 – ident: 139_CR32 – ident: 139_CR36 – ident: 139_CR63 doi: 10.1609/aaai.v34i01.5467 – ident: 139_CR37 doi: 10.1109/TITS.2023.3263847 – volume-title: Pattern recognition and machine learning year: 2006 ident: 139_CR5 – volume: 21 start-page: 1086 issue: 3 year: 2019 ident: 139_CR8 publication-title: IEEE Trans Intell Transp Syst doi: 10.1109/TITS.2019.2901791 – ident: 139_CR15 – volume: 130 year: 2021 ident: 139_CR62 publication-title: Transp Res C Emerg Technol doi: 10.1016/j.trc.2021.103321 – ident: 139_CR22 – volume: 68 start-page: 1243 issue: 2 year: 2019 ident: 139_CR30 publication-title: IEEE Trans Veh Technol doi: 10.1109/TVT.2018.2890726 – ident: 139_CR19 – ident: 139_CR68 – volume: 4 start-page: 42 issue: 1 year: 1956 ident: 139_CR46 publication-title: Oper Res doi: 10.1287/opre.4.1.42 – volume: 129 start-page: 278 issue: 3 year: 2003 ident: 139_CR1 publication-title: J Transp Eng doi: 10.1061/(ASCE)0733-947X(2003)129:3(278) – volume: 33 start-page: 4079 year: 2020 ident: 139_CR43 publication-title: Adv Neural Inf Process Syst – ident: 139_CR72 doi: 10.1007/s42421-020-00029-6
SSID	ssj0002901294
Score	2.3106189
Snippet	The optimization of traffic signal control (TSC) is critical for an efficient transportation system. In recent years, reinforcement learning (RL) techniques...
SourceID	proquest crossref springer
SourceType	Aggregation Database Index Database Publisher
StartPage	25
SubjectTerms	Adaptive control Computational Intelligence Data Mining and Knowledge Discovery Datasets Engineering Flow theory Machine learning Neural networks Optimization Regularization methods Simulation Simulators Traffic control Traffic flow Traffic information Traffic intersections Traffic signals Transportation systems Transportation Technology and Traffic Engineering
Title	A Fully Data-Driven Approach for Realistic Traffic Signal Control Using Offline Reinforcement Learning
URI	https://link.springer.com/article/10.1007/s42421-025-00139-z https://www.proquest.com/docview/3268302929
Volume	7
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVAVX databaseName: SpringerLINK Contemporary 1997-Present customDbUrl: eissn: 2948-1368 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002901294 issn: 2948-135X databaseCode: RSV dateStart: 20190601 isFulltext: true titleUrlDefault: https://link.springer.com/search?facet-content-type=%22Journal%22 providerName: Springer Nature
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV1NTwIxEJ0oetCD30YUTQ_etMlut2y7RwISDwYNKOG2abst4YIG0ER-vW3pgho96GUv20yamU7ntTPzCnAZm4xFmVTYkEJhyrX1ubpIcaZdnyaNeOGvBvp3rNPhg0H2EJrCpmW1e5mS9Dv1stmNuuwlds-vetyC5-uwYcMdd-7Y7fWXNysuM0j8C4j2a49ISX0QumV-FvM1Iq1g5rfMqA847d3_TXUPdgLARI3FitiHNT0-gO1PtIOHYBrInTzfUUvMBG5N3IaHGoFcHFkUi7ra8SJaCcjGMkcygXqjoRPbXFS2I19pgO6NcSjVDvf8q8pfNaJA2To8gqf2zWPzFof3FrCKOZtjwgWjkUwSoaioZyTiJrI2lgVPiogaxSktWGyoFiRNdCYoN4QwQ5QhXMpUJsdQGT-P9QkgJhkVkjHN04gKTrkUFhc4djiL_wSJq3BV6jx_WdBq5EsCZa-93Gov99rL51WolWbJg4tNc4s7HXeZhXdVuC7NsPr9u7TTvw0_gy3iLOlLWGpQmU1e9TlsqrfZaDq58EvvA_yB0yI
linkProvider	Springer Nature
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV3PT8IwFH5RNFEP_jaiqD140yZb19HuSECCEdEAEm5Lu7WECxpAE_nrbcsGavSglx225mV5r-372vfe9wAufR0xL5IJ1iRNMOXKrLlQlHGkbJ0m9XjqrgZ6TdZq8X4_esyKwiZ5tnseknQ79aLYjdroJbbtVx1uwbNVWDNvQpvI1-70FjcrNjJIXAdE8zRHpCDsZ9UyP4v56pGWMPNbZNQ5nPrO_351F7YzgIkq8xmxBytqtA9bn2gHD0BXkD15vqOamApcG9sND1UycnFkUCxqK8uLaCQg48ssyQTqDAdWbHWe2Y5cpgF60NqiVDPc8a8m7qoRZZStg0N4qt90qw2c9VvAic_ZDBMuGPVkEIiEijAiHteesbFMeZB6VCec0pT5mipByoGKBOWaEKZJogmXsiyDIyiMnkfqGBCTjArJmOJljwpOuRQGF1h2OIP_BPGLcJXrPH6Z02rECwJlp73YaC922otnRSjlZomzJTaJDe603GUG3hXhOjfD8vPv0k7-NvwCNhrd-2bcvG3dncImsVZ16SwlKEzHr-oM1pO36XAyPnfT8APrptYG
linkToPdf	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LT8MwDLZgIAQH3ojBgBy4QUSbZk16nDYmENOYGEy7VUmbTLuMaRtI7NeTpN0DBAfEpZdGVmUn9efY_gxw6euIeZFMsCZpgilX5syVRYgjZfs0qcdTdzXQabBmk3e7UWupi99Vu89SkllPg2VpGkxuhqm-mTe-UZvJxHYUq8MweLoKa9QODbLxerszv2WxWULipiGapwmXgnI375z5WcxX77SAnN-ypM751Hf-_9m7sJ0DT1TJdsoerKjBPmwt0REegK4gG5F-oJqYCFwb2R8hquSk48igW_SkLF-ikYCMj7PkE6jd71mx1aziHbkKBPSotUWvZrnjZU3cFSTKqVx7h_BSv32u3uF8DgNOfM6mmHDBqCeDQCRUlCPice0Z28uUB6lHdcIpTZmvqRIkDFQkKNeEME0STbiUoQyOoDB4HahjQEwyKiRjioceFZxyKQxesKxxBhcK4hfhaqb_eJjRbcRzYmWnvdhoL3bai6dFKM1MFOdHbxwbPGo5zQzsK8L1zCSL179LO_nb8gvYaNXqceO--XAKm8Qa1VW5lKAwGb2pM1hP3if98ejc7chPMW_e6g
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=A+Fully+Data-Driven+Approach+for+Realistic+Traffic+Signal+Control+Using+Offline+Reinforcement+Learning&rft.jtitle=Data+science+for+Transportation&rft.au=Li%2C+Jianxiong&rft.au=Lin%2C+Shichao&rft.au=Shi%2C+Tianyu&rft.au=Tian%2C+Chujie&rft.date=2025-12-01&rft.issn=2948-135X&rft.eissn=2948-1368&rft.volume=7&rft.issue=3&rft_id=info:doi/10.1007%2Fs42421-025-00139-z&rft.externalDBID=n%2Fa&rft.externalDocID=10_1007_s42421_025_00139_z
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2948-135X&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2948-135X&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2948-135X&client=summon