Automated Configuration of Evolutionary Algorithms via Deep Reinforcement Learning for Constrained Multiobjective Optimization

Learning to optimize and automated algorithm design are attracting increasing attention, but it is still in its infancy in constrained multiobjective optimization evolutionary algorithms (CMOEAs). Current learning-assisted CMOEAs are typically crafted by human experts using manually designed techniq...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on cybernetics Vol. 55; no. 12; pp. 1 - 14
Main Authors:	Ming, Fei, Gong, Wenyin, Xue, Bing, Zhang, Mengjie, Jin, Yaochu
Format:	Journal Article
Language:	English
Published:	United States IEEE 01.12.2025
Subjects:	Automated algorithm configuration constrained multiobjective optimization Convergence Cybernetics Deep reinforcement learning DRL evolutionary algorithms (EAs) Evolutionary computation Evolutionary dynamics Heuristic algorithms learning to optimize Linear programming Optimization Training Vectors
ISSN:	2168-2267, 2168-2275
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	Learning to optimize and automated algorithm design are attracting increasing attention, but it is still in its infancy in constrained multiobjective optimization evolutionary algorithms (CMOEAs). Current learning-assisted CMOEAs are typically crafted by human experts using manually designed techniques, which tend to be overly tuned, ad hoc, and lacking versatility. To alleviate these limitations, this work proposes transforming the online configuration of CMOEA into determinations of discrete and continuous parameters, which are then solved by deep reinforcement learning (DRL) techniques. Specifically, the Actor-Critic framework is adapted to determine a factor that defines the environmental selection pressure. The deep Q -learning technique is adopted to determine the operators for producing offspring. Owing to the property of DRL, the configured algorithm can accommodate historical experience, current evolutionary dynamics, and future improvements to achieve self-learning. A new CMOEA is proposed using the automatically configured EA. Experiments on four challenging benchmarks and 21 real-world problems verify that our method significantly outperforms 11 state-of-the-art methods. The versatility and superiority of the automatically configured environment and operators over handcrafted methods justify the effectiveness of the automated configuration method, demonstrating a promising direction in evolutionary multiobjective optimization.
AbstractList	Learning to optimize and automated algorithm design are attracting increasing attention, but it is still in its infancy in constrained multiobjective optimization evolutionary algorithms (CMOEAs). Current learning-assisted CMOEAs are typically crafted by human experts using manually designed techniques, which tend to be overly tuned, ad hoc, and lacking versatility. To alleviate these limitations, this work proposes transforming the online configuration of CMOEA into determinations of discrete and continuous parameters, which are then solved by deep reinforcement learning (DRL) techniques. Specifically, the Actor-Critic framework is adapted to determine a factor that defines the environmental selection pressure. The deep Q -learning technique is adopted to determine the operators for producing offspring. Owing to the property of DRL, the configured algorithm can accommodate historical experience, current evolutionary dynamics, and future improvements to achieve self-learning. A new CMOEA is proposed using the automatically configured EA. Experiments on four challenging benchmarks and 21 real-world problems verify that our method significantly outperforms 11 state-of-the-art methods. The versatility and superiority of the automatically configured environment and operators over handcrafted methods justify the effectiveness of the automated configuration method, demonstrating a promising direction in evolutionary multiobjective optimization. Learning to optimize and automated algorithm design are attracting increasing attention, but it is still in its infancy in constrained multiobjective optimization evolutionary algorithms (CMOEAs). Current learning-assisted CMOEAs are typically crafted by human experts using manually designed techniques, which tend to be overly tuned, ad hoc, and lacking versatility. To alleviate these limitations, this work proposes transforming the online configuration of CMOEA into determinations of discrete and continuous parameters, which are then solved by deep reinforcement learning (DRL) techniques. Specifically, the Actor-Critic framework is adapted to determine a factor that defines the environmental selection pressure. The deep Q-learning technique is adopted to determine the operators for producing offspring. Owing to the property of DRL, the configured algorithm can accommodate historical experience, current evolutionary dynamics, and future improvements to achieve self-learning. A new CMOEA is proposed using the automatically configured evolutionary algorithm. Experiments on four challenging benchmarks and 21 real-world problems verify that our method significantly outperforms 11 state-of-the-art methods. The versatility and superiority of the automatically configured environment and operators over handcrafted methods justify the effectiveness of the automated configuration method, demonstrating a promising direction in evolutionary multiobjective optimization.
Author	Xue, Bing Zhang, Mengjie Gong, Wenyin Jin, Yaochu Ming, Fei
Author_xml	– sequence: 1 givenname: Fei orcidid: 0000-0003-1042-8206 surname: Ming fullname: Ming, Fei email: feiming@cug.edu.cn organization: School of Engineering, Trustworthy and General Artificial Intelligence Laboratory, Westlake University, Hangzhou, China – sequence: 2 givenname: Wenyin orcidid: 0000-0003-1610-6865 surname: Gong fullname: Gong, Wenyin email: wygong@cug.edu.cn organization: School of Computer Science, China University of Geosciences, Wuhan, China – sequence: 3 givenname: Bing orcidid: 0000-0002-4865-8026 surname: Xue fullname: Xue, Bing email: bing.xue@ecs.vuw.ac.nz organization: School of Engineering and Computer Science, Victoria University of Wellington, Wellington, New Zealand – sequence: 4 givenname: Mengjie orcidid: 0000-0003-4463-9538 surname: Zhang fullname: Zhang, Mengjie email: mengjie.zhang@ecs.vuw.ac.nz organization: School of Engineering and Computer Science, Victoria University of Wellington, Wellington, New Zealand – sequence: 5 givenname: Yaochu orcidid: 0000-0003-1100-0631 surname: Jin fullname: Jin, Yaochu email: jinyaochu@westlake.edu.cn organization: School of Engineering, Trustworthy and General Artificial Intelligence Laboratory, Westlake University, Hangzhou, China
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/40911451$$D View this record in MEDLINE/PubMed
BookMark	eNpFkNtKw0AQhhepWK19AEFkXyB1D8k2uay1HqBSkHrhVdhsJnVLshs2m4Je-Owmtta5mQP__MN8F2hgrAGEriiZUEqS2_X8_W7CCIsmXBDOInqCzhkVccDYNBocazEdonHTbEkXcTdK4jM0DElCaRjRc_Q9a72tpIccz60p9KZ10mtrsC3wYmfLtm-k-8SzcmOd9h9Vg3da4nuAGr-CNoV1CiowHi9BOqPNBnej3qzxTmrTGb-0ZeeSbUF5vQO8qr2u9NfvmUt0WsiygfEhj9Dbw2I9fwqWq8fn-WwZKBZyH8iYx0mYMKWKsMgjyCXhnBClRCaKaR6ROJJZnolQCimJkKGQSaioglxAlrGMj9DN3rduswrytHa66r5K_0B0AroXKGebxkFxlFCS9rzTnnfa804PvLud6_2OBoB_PaUR44zyHy3If3o
CODEN	ITCEB8
Cites_doi	10.1038/nature14236 10.1109/TEVC.2023.3292527 10.1016/j.ins.2017.10.037 10.1109/MCI.2023.3245719 10.1016/j.eswa.2021.115493 10.23919/cje.2022.00.100 10.1109/tevc.2024.3395699 10.1109/TCYB.2023.3341982 10.1007/978-3-030-12598-1_27 10.1109/TCYB.2022.3178132 10.1016/j.ins.2014.03.042 10.1109/TEVC.2021.3066301 10.1109/TEVC.2023.3260306 10.1109/TCYB.2022.3163759 10.1109/TETCI.2024.3393368 10.1109/tevc.2024.3525153 10.1109/tevc.2024.3416552 10.1016/j.swevo.2022.101178 10.1109/MCI.2017.2742868 10.1109/TEVC.2022.3194729 10.1109/TCYB.2020.3021138 10.1109/TEVC.2018.2855411 10.1145/3638529.3653994 10.1109/TEVC.2022.3175065 10.1109/TEVC.2020.3004012 10.1162/evco_a_00259 10.1109/TVT.2025.3553499 10.1016/j.neucom.2024.129296 10.1016/j.swevo.2018.08.017 10.1016/j.swevo.2021.100961 10.1109/TEVC.2019.2894743 10.1109/TEVC.2023.3345470 10.1109/TEVC.2018.2869001 10.1109/TEVC.2022.3224600 10.1109/TEVC.2022.3155533 10.1016/j.engappai.2024.108673 10.1109/TEVC.2023.3328886 10.1109/TSMC.2019.2943973 10.1109/TEVC.2022.3197298 10.1109/TCYB.2021.3089633 10.1109/tevc.2025.3558362 10.1109/4235.797969 10.1109/TKDE.2009.191 10.1109/TEVC.2013.2281534 10.1109/TCYB.2023.3329947 10.1145/3520304.3533640 10.1109/TEVC.2022.3199775 10.1007/s00500-019-03794-x 10.1145/1390156.1390294 10.1109/TEVC.2008.925798 10.1007/978-3-030-58112-1_48 10.1109/TEVC.2022.3186667 10.1109/TPAMI.2025.3554669 10.1109/TEVC.2021.3060811 10.1109/4235.996017 10.1613/jair.1.13676 10.1145/3321707.3321813 10.1109/TC.2022.3173151 10.1109/MCI.2023.3277768 10.1016/j.ejor.2021.05.042
ContentType	Journal Article
DBID	97E RIA RIE AAYXX CITATION NPM
DOI	10.1109/TCYB.2025.3603251
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed
DatabaseTitle	CrossRef PubMed
DatabaseTitleList	PubMed
Database_xml	– sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Sciences (General)
EISSN	2168-2275
EndPage	14
ExternalDocumentID	40911451 10_1109_TCYB_2025_3603251 11152321
Genre	orig-research Journal Article
GrantInformation_xml	– fundername: National Natural Science Foundation of China grantid: 62576325 funderid: 10.13039/501100001809 – fundername: International Collaboration Fund for Creative Research of National Science Foundation of China (NSFC ICFCRT) grantid: W2441019
GroupedDBID	0R~ 4.4 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK AENEX AGQYO AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ EBS HZ~ IFIPE IPLJI JAVBF M43 O9- OCL PQQKQ RIA RIE RNS AAYXX AGSQL CITATION EJD NPM
ID	FETCH-LOGICAL-c243t-a8389492ccf4fd5eda03300cc6b6f7d5085abdb64a6aa06a46a94c1ced6ebb2b3
IEDL.DBID	RIE
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001566853600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	2168-2267
IngestDate	Fri Nov 28 01:41:34 EST 2025 Thu Nov 27 00:55:19 EST 2025 Wed Sep 17 06:32:00 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	12
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c243t-a8389492ccf4fd5eda03300cc6b6f7d5085abdb64a6aa06a46a94c1ced6ebb2b3
ORCID	0000-0003-1042-8206 0000-0003-4463-9538 0000-0003-1610-6865 0000-0003-1100-0631 0000-0002-4865-8026
PMID	40911451
PageCount	14
ParticipantIDs	crossref_primary_10_1109_TCYB_2025_3603251 pubmed_primary_40911451 ieee_primary_11152321
PublicationCentury	2000
PublicationDate	2025-Dec
PublicationDateYYYYMMDD	2025-12-01
PublicationDate_xml	– month: 12 year: 2025 text: 2025-Dec
PublicationDecade	2020
PublicationPlace	United States
PublicationPlace_xml	– name: United States
PublicationTitle	IEEE transactions on cybernetics
PublicationTitleAbbrev	TCYB
PublicationTitleAlternate	IEEE Trans Cybern
PublicationYear	2025
Publisher	IEEE
Publisher_xml	– name: IEEE
References	ref13 ref57 ref12 ref15 ref59 ref14 ref53 ref52 ref11 ref55 ref10 ref17 ref16 ref19 ref18 Zhao (ref20) 2023 ref51 ref50 Zhang (ref60) 2008 ref45 ref48 Deb (ref58) 1995; 9 ref42 ref41 ref44 ref49 ref8 ref7 ref9 ref4 ref3 ref6 ref5 ref40 ref35 ref34 ref37 ref36 ref31 ref30 ref33 ref32 ref2 ref1 ref39 Adriaensen (ref47) ref38 Zitzler (ref56) ref24 ref68 ref23 ref67 ref26 ref25 ref64 ref63 ref22 ref66 ref21 ref65 ref28 ref27 Xue (ref43); 35 ref29 Biedenkapp (ref46) Silver (ref54) ref62 ref61
References_xml	– ident: ref18 doi: 10.1038/nature14236 – ident: ref2 doi: 10.1109/TEVC.2023.3292527 – ident: ref26 doi: 10.1016/j.ins.2017.10.037 – ident: ref39 doi: 10.1109/MCI.2023.3245719 – ident: ref48 doi: 10.1016/j.eswa.2021.115493 – ident: ref42 doi: 10.23919/cje.2022.00.100 – ident: ref53 doi: 10.1109/tevc.2024.3395699 – ident: ref5 doi: 10.1109/TCYB.2023.3341982 – ident: ref65 doi: 10.1007/978-3-030-12598-1_27 – ident: ref38 doi: 10.1109/TCYB.2022.3178132 – ident: ref17 doi: 10.1016/j.ins.2014.03.042 – ident: ref24 doi: 10.1109/TEVC.2021.3066301 – start-page: 387 volume-title: Proc. 31st Int. Conf. Mach. Learn. ident: ref54 article-title: Deterministic policy gradient algorithms – ident: ref36 doi: 10.1109/TEVC.2023.3260306 – ident: ref11 doi: 10.1109/TCYB.2022.3163759 – ident: ref37 doi: 10.1109/TETCI.2024.3393368 – volume-title: Multiobjective Optimization Test Instances for the CEC 2009 Special Session and Competition year: 2008 ident: ref60 – ident: ref27 doi: 10.1109/tevc.2024.3525153 – ident: ref35 doi: 10.1109/tevc.2024.3416552 – ident: ref28 doi: 10.1016/j.swevo.2022.101178 – ident: ref59 doi: 10.1109/MCI.2017.2742868 – ident: ref63 doi: 10.1109/TEVC.2022.3194729 – year: 2023 ident: ref20 article-title: Automated design of metaheuristic algorithms: A survey publication-title: arXiv:2303.06532 – ident: ref14 doi: 10.1109/TCYB.2020.3021138 – ident: ref15 doi: 10.1109/TEVC.2018.2855411 – ident: ref8 doi: 10.1145/3638529.3653994 – ident: ref10 doi: 10.1109/TEVC.2022.3175065 – ident: ref16 doi: 10.1109/TEVC.2020.3004012 – start-page: 742 volume-title: Proc. Evol. Methods Design Optim. Control Appl. Ind. Probl. ident: ref56 article-title: SPEA2+: Improving the strength pareto evolutionary algorithm – ident: ref61 doi: 10.1162/evco_a_00259 – ident: ref6 doi: 10.1109/TVT.2025.3553499 – start-page: 554 volume-title: Proc. 25th Int. Joint Conf. Artif. Intell. ident: ref47 article-title: Towards a white box approach to automated algorithm design – ident: ref68 doi: 10.1016/j.neucom.2024.129296 – ident: ref13 doi: 10.1016/j.swevo.2018.08.017 – ident: ref3 doi: 10.1016/j.swevo.2021.100961 – ident: ref30 doi: 10.1109/TEVC.2019.2894743 – ident: ref34 doi: 10.1109/TEVC.2023.3345470 – ident: ref21 doi: 10.1109/TEVC.2018.2869001 – ident: ref31 doi: 10.1109/TEVC.2022.3224600 – ident: ref7 doi: 10.1109/TEVC.2022.3155533 – ident: ref9 doi: 10.1016/j.engappai.2024.108673 – ident: ref4 doi: 10.1109/TEVC.2023.3328886 – ident: ref64 doi: 10.1109/TSMC.2019.2943973 – ident: ref52 doi: 10.1109/TEVC.2022.3197298 – ident: ref40 doi: 10.1109/TCYB.2021.3089633 – ident: ref33 doi: 10.1109/tevc.2025.3558362 – volume: 9 start-page: 115 issue: 2 year: 1995 ident: ref58 article-title: Simulated binary crossover for continuous search space publication-title: Complex Syst. – volume: 35 start-page: 20147 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref43 article-title: Multi-agent dynamic algorithm configuration – start-page: 427 volume-title: Proc. ECAI ident: ref46 article-title: Dynamic algorithm configuration: Foundation of a new meta-algorithmic framework – ident: ref66 doi: 10.1109/4235.797969 – ident: ref67 doi: 10.1109/TKDE.2009.191 – ident: ref62 doi: 10.1109/TEVC.2013.2281534 – ident: ref32 doi: 10.1109/TCYB.2023.3329947 – ident: ref23 doi: 10.1145/3520304.3533640 – ident: ref25 doi: 10.1109/TEVC.2022.3199775 – ident: ref29 doi: 10.1007/s00500-019-03794-x – ident: ref55 doi: 10.1145/1390156.1390294 – ident: ref57 doi: 10.1109/TEVC.2008.925798 – ident: ref50 doi: 10.1007/978-3-030-58112-1_48 – ident: ref41 doi: 10.1109/TEVC.2022.3186667 – ident: ref45 doi: 10.1109/TPAMI.2025.3554669 – ident: ref51 doi: 10.1109/TEVC.2021.3060811 – ident: ref22 doi: 10.1109/4235.996017 – ident: ref44 doi: 10.1613/jair.1.13676 – ident: ref49 doi: 10.1145/3321707.3321813 – ident: ref1 doi: 10.1109/TC.2022.3173151 – ident: ref12 doi: 10.1109/MCI.2023.3277768 – ident: ref19 doi: 10.1016/j.ejor.2021.05.042
SSID	ssj0000816898
Score	2.4120598
Snippet	Learning to optimize and automated algorithm design are attracting increasing attention, but it is still in its infancy in constrained multiobjective...
SourceID	pubmed crossref ieee
SourceType	Index Database Publisher
StartPage	1
SubjectTerms	Automated algorithm configuration constrained multiobjective optimization Convergence Cybernetics Deep reinforcement learning DRL evolutionary algorithms (EAs) Evolutionary computation Evolutionary dynamics Heuristic algorithms learning to optimize Linear programming Optimization Training Vectors
Title	Automated Configuration of Evolutionary Algorithms via Deep Reinforcement Learning for Constrained Multiobjective Optimization
URI	https://ieeexplore.ieee.org/document/11152321 https://www.ncbi.nlm.nih.gov/pubmed/40911451
Volume	55
WOSCitedRecordID	wos001566853600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 2168-2275 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000816898 issn: 2168-2267 databaseCode: RIE dateStart: 20130101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV05T8MwGLWgYmABCgXKJQ8MgBRwYsdJxnKJARWEQCpT5ONLKYIGtWklFn47tmOKQGJgi3JYkZ-P99l-30Nov1Cplvb0jMoYCVhCZSDMvBfYHSWSyowWwrmWXCfdbtrrZbderO60MADgDp_Bsb10e_m6VBO7VHZi-qWJm6xsfD5JklqsNVtQcQ4Szvs2MheBoRWJ38UMSXZyf_Z4aqLBKD6mnNAoDn_MQ85Y5RexdBPM5fI_f20FLXkmiTs19E00B8NV1PR9dYwPfELpwzX00ZlUpWGmoLEV-A36kxp2XBb4Yurbnhi9485LvxwNqqfXMZ4OBD4HeMN34HKrKreMiH061j42t2xhY2cxYQp2Qt5SPtfjJ74xI9Grl3i20MPlxf3ZVeB9FwIVMVoFIjUshmWRUgUrdAxaEEoJUYpLXiTaULpYSC05E1wIwgXjImMqVKA5SBlJuo4aw3IImwgrzoFmJJbaEC_DrtICBMhCS2DAQ5BtdPSFQv5Wp9fIXVhCstxCllvIcg9ZG7VsxX-_6Ou8jTZq4GZPTNAaWgvirT--2EaLtuD6WMoOalSjCeyiBTWtBuPRnmtXn3UVzjw
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3JTsMwELUQIMGFfSmrDxwAKeDEjpscyyYQpSBUJDhFXialCBrUppW48O3YjgGBxIGblcWJ_Ly8GfvNILSTq0RLe3pGpYwErE5lIMy6F9gdJZLIlObCZS1p1lut5P4-vfFidaeFAQB3-AwObNHt5etCDa2r7NCMS2M3Wdn4RMxYFFZyrS-Xissh4bLfRqYQGGJR9_uYIUkP28cPR8YejOIDygmN4vDHSuRSq_yilm6JOZv958_NoRnPJXGjAn8ejUFvAc370TrAuz6k9N4iem8My8JwU9DYSvy6nWEFPC5yfDryvU_033DjuVP0u-XjywCPugKfALziW3DRVZVzJGIfkLWDzSVb2cAlmTAVOylvIZ-qGRRfm7noxYs8l9Dd2Wn7-DzwmRcCFTFaBiIxPIalkVI5y3UMWhBKCVGKS57XtSF1sZBacia4EIQLxkXKVKhAc5AyknQZjfeKHqwirDgHmpJYakO9DL9KchAgcy2BAQ9B1tD-JwrZaxVgI3OGCUkzC1lmIcs8ZDW0ZBv--0Hf5jW0UgH3dceYraFNQrz2xxvbaOq8fdXMmhety3U0bT9SHVLZQONlfwibaFKNyu6gv-X62AdUftGD
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=Automated+Configuration+of+Evolutionary+Algorithms+via+Deep+Reinforcement+Learning+for+Constrained+Multiobjective+Optimization&rft.jtitle=IEEE+transactions+on+cybernetics&rft.au=Ming%2C+Fei&rft.au=Gong%2C+Wenyin&rft.au=Xue%2C+Bing&rft.au=Zhang%2C+Mengjie&rft.date=2025-12-01&rft.eissn=2168-2275&rft.volume=55&rft.issue=12&rft.spage=5877&rft_id=info:doi/10.1109%2FTCYB.2025.3603251&rft_id=info%3Apmid%2F40911451&rft.externalDocID=40911451
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2168-2267&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2168-2267&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2168-2267&client=summon