Learning Feature Embedding Refiner for Solving Vehicle Routing Problems
While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are less effective in searching solutions due to deterministic feature embeddings and deterministic probability distributions. In this article, we...
Gespeichert in:
| Veröffentlicht in: | IEEE transaction on neural networks and learning systems Jg. 35; H. 11; S. 15279 - 15291 |
|---|---|
| Hauptverfasser: | , , , , , , |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
United States
IEEE
01.11.2024
|
| Schlagworte: | |
| ISSN: | 2162-237X, 2162-2388, 2162-2388 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Abstract | While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are less effective in searching solutions due to deterministic feature embeddings and deterministic probability distributions. In this article, we propose the feature embedding refiner (FER) with a novel and generic encoder-refiner-decoder structure to boost the existing encoder-decoder structured deep models. It is model-agnostic that the encoder and the decoder can be from any pretrained neural construction method. Regarding the introduced refiner network, we design its architecture by combining the standard gated recurrent units (GRU) cell with two new layers, i.e., an accumulated graph attention (AGA) layer and a gated nonlinear (GNL) layer. The former extracts dynamic graph topological information of historical solutions stored in a diversified solution pool to generate aggregated pool embeddings that are further improved by the GRU, and the latter adaptively refines the feature embeddings from the encoder with the guidance of the improved pool embeddings. To this end, our FER allows current neural construction methods to not only iteratively refine the feature embeddings for boarder search range but also dynamically update the probability distributions for more diverse search. We apply FER to two prevailing neural construction methods including attention model (AM) and policy optimization with multiple optima (POMO) to solve the traveling salesman problem (TSP) and the capacitated VRP (CVRP). Experimental results show that our method achieves lower gaps and better generalization than the original ones and also exhibits competitive performance to the state-of-the-art neural improvement methods. |
|---|---|
| AbstractList | While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are less effective in searching solutions due to deterministic feature embeddings and deterministic probability distributions. In this article, we propose the feature embedding refiner (FER) with a novel and generic encoder-refiner-decoder structure to boost the existing encoder-decoder structured deep models. It is model-agnostic that the encoder and the decoder can be from any pretrained neural construction method. Regarding the introduced refiner network, we design its architecture by combining the standard gated recurrent units (GRU) cell with two new layers, i.e., an accumulated graph attention (AGA) layer and a gated nonlinear (GNL) layer. The former extracts dynamic graph topological information of historical solutions stored in a diversified solution pool to generate aggregated pool embeddings that are further improved by the GRU, and the latter adaptively refines the feature embeddings from the encoder with the guidance of the improved pool embeddings. To this end, our FER allows current neural construction methods to not only iteratively refine the feature embeddings for boarder search range but also dynamically update the probability distributions for more diverse search. We apply FER to two prevailing neural construction methods including attention model (AM) and policy optimization with multiple optima (POMO) to solve the traveling salesman problem (TSP) and the capacitated VRP (CVRP). Experimental results show that our method achieves lower gaps and better generalization than the original ones and also exhibits competitive performance to the state-of-the-art neural improvement methods. While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are less effective in searching solutions due to deterministic feature embeddings and deterministic probability distributions. In this article, we propose the feature embedding refiner (FER) with a novel and generic encoder-refiner-decoder structure to boost the existing encoder-decoder structured deep models. It is model-agnostic that the encoder and the decoder can be from any pretrained neural construction method. Regarding the introduced refiner network, we design its architecture by combining the standard gated recurrent units (GRU) cell with two new layers, i.e., an accumulated graph attention (AGA) layer and a gated nonlinear (GNL) layer. The former extracts dynamic graph topological information of historical solutions stored in a diversified solution pool to generate aggregated pool embeddings that are further improved by the GRU, and the latter adaptively refines the feature embeddings from the encoder with the guidance of the improved pool embeddings. To this end, our FER allows current neural construction methods to not only iteratively refine the feature embeddings for boarder search range but also dynamically update the probability distributions for more diverse search. We apply FER to two prevailing neural construction methods including attention model (AM) and policy optimization with multiple optima (POMO) to solve the traveling salesman problem (TSP) and the capacitated VRP (CVRP). Experimental results show that our method achieves lower gaps and better generalization than the original ones and also exhibits competitive performance to the state-of-the-art neural improvement methods.While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are less effective in searching solutions due to deterministic feature embeddings and deterministic probability distributions. In this article, we propose the feature embedding refiner (FER) with a novel and generic encoder-refiner-decoder structure to boost the existing encoder-decoder structured deep models. It is model-agnostic that the encoder and the decoder can be from any pretrained neural construction method. Regarding the introduced refiner network, we design its architecture by combining the standard gated recurrent units (GRU) cell with two new layers, i.e., an accumulated graph attention (AGA) layer and a gated nonlinear (GNL) layer. The former extracts dynamic graph topological information of historical solutions stored in a diversified solution pool to generate aggregated pool embeddings that are further improved by the GRU, and the latter adaptively refines the feature embeddings from the encoder with the guidance of the improved pool embeddings. To this end, our FER allows current neural construction methods to not only iteratively refine the feature embeddings for boarder search range but also dynamically update the probability distributions for more diverse search. We apply FER to two prevailing neural construction methods including attention model (AM) and policy optimization with multiple optima (POMO) to solve the traveling salesman problem (TSP) and the capacitated VRP (CVRP). Experimental results show that our method achieves lower gaps and better generalization than the original ones and also exhibits competitive performance to the state-of-the-art neural improvement methods. |
| Author | Li, Jingwen Cao, Zhiguang Ma, Yining Wu, Yaoxin Song, Wen Zhang, Jie Chee, Yeow Meng |
| Author_xml | – sequence: 1 givenname: Jingwen orcidid: 0000-0001-8661-5695 surname: Li fullname: Li, Jingwen email: lijingwen@sicnu.edu.cn organization: Department of Computer Science, Sichuan Normal University, Chengdu, China – sequence: 2 givenname: Yining orcidid: 0000-0002-6639-8547 surname: Ma fullname: Ma, Yining email: yiningma@u.nus.edu organization: Department of Industrial Systems Engineering and Management, College of Design and Engineering, National University of Singapore, Queenstown, Singapore – sequence: 3 givenname: Zhiguang orcidid: 0000-0002-4499-759X surname: Cao fullname: Cao, Zhiguang email: zhiguangcao@outlook.com organization: School of Computing and Information Systems, Singapore Management University, Bras Basah, Singapore – sequence: 4 givenname: Yaoxin orcidid: 0000-0002-3625-6599 surname: Wu fullname: Wu, Yaoxin email: wyxacc@hotmail.com organization: Faculty of Industrial Engineering and Innovation Sciences, Eindhoven University of Technology, Eindhoven, The Netherlands – sequence: 5 givenname: Wen orcidid: 0000-0001-7624-1861 surname: Song fullname: Song, Wen email: wensong@email.sdu.edu.cn organization: Institute of Marine Science and Technology, Shandong University, Qingdao, China – sequence: 6 givenname: Jie surname: Zhang fullname: Zhang, Jie email: zhangj@ntu.edu.sg organization: School of Computer Science and Engineering, Nanyang Technological University, Jurong West, Singapore – sequence: 7 givenname: Yeow Meng surname: Chee fullname: Chee, Yeow Meng email: ymchee@nus.edu.sg organization: Department of Industrial Systems Engineering and Management, College of Design and Engineering, National University of Singapore, Queenstown, Singapore |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/37352084$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9kF1LwzAUhoMofv8BEemlN5vJSdeklyLbFIaKTvEupOmJRtpGk1bw37u6TcQLz8354H1eOO8e2Wx8g4QcMTpkjOZn8-vr2f0QKPAhBzmiQmyQXWAZDIBLufkzi6cdchjjK11URkdZmm-THS74CKhMd8l0hjo0rnlOJqjbLmAyrgssy_5yh9Y1GBLrQ3Lvq4_-9ogvzlSY3Pmu7ffb4IsK63hAtqyuIh6u-j55mIznF5eD2c306uJ8NjA8k-1AaJozkBbAoi20RC1QQikMaJNzY1Nrgcm0TIsyLwtRWmYMFxkXBWVCGMn3yenS9y349w5jq2oXDVaVbtB3UYGEPAUhGF9IT1bSrqixVG_B1Tp8qvXzC4FcCkzwMQa0yrhWt843bdCuUoyqPmr1HbXqo1arqBco_EHX7v9Cx0vIIeIvgGWUpiP-BSG2ieU |
| CODEN | ITNNAL |
| CitedBy_id | crossref_primary_10_1109_TNNLS_2024_3371781 crossref_primary_10_3390_math13132039 crossref_primary_10_1007_s40747_025_01826_8 crossref_primary_10_1177_03611981251322486 crossref_primary_10_3390_app15094776 crossref_primary_10_1007_s40747_025_02018_0 crossref_primary_10_1007_s10489_025_06778_7 |
| Cites_doi | 10.1016/j.ejor.2016.08.012 10.1109/TNNLS.2017.2740224 10.1016/j.cor.2012.07.018 10.1007/s10479-009-0657-6 10.1109/TSMC.2020.2969317 10.1016/j.cor.2021.105643 10.1016/j.ejor.2004.08.018 10.1109/TNNLS.2021.3068828 10.1007/978-3-319-60801-3_27 10.1109/tnnls.2022.3148435 10.1287/ijoc.3.4.376 10.1109/CVPR.2016.90 10.1109/TITS.2021.3056120 10.5555/3045118.3045167 10.1109/TII.2020.3031409 10.1109/TNNLS.2021.3105905 10.1016/j.eswa.2011.03.088 10.1016/0377-2217(92)90192-C 10.1016/S1352-2310(97)00447-0 10.1109/TNNLS.2021.3070584 10.48550/ARXIV.1706.03762 |
| ContentType | Journal Article |
| DBID | 97E RIA RIE AAYXX CITATION NPM 7X8 |
| DOI | 10.1109/TNNLS.2023.3285077 |
| DatabaseName | IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed MEDLINE - Academic |
| DatabaseTitle | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic 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 – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| EISSN | 2162-2388 |
| EndPage | 15291 |
| ExternalDocumentID | 37352084 10_1109_TNNLS_2023_3285077 10160045 |
| Genre | orig-research Journal Article |
| GrantInformation_xml | – fundername: Natural Science Foundation of Shandong Province grantid: ZR2021QF063 funderid: 10.13039/501100007129 – fundername: National Natural Science Foundation of China grantid: 62102228 funderid: 10.13039/501100001809 – fundername: Agency for Science Technology and Research Career Development Fund grantid: C222812027 |
| GroupedDBID | 0R~ 4.4 5VS 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK ACPRK AENEX AFRAH AGQYO AGSQL AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ EBS EJD IFIPE IPLJI JAVBF M43 MS~ O9- OCL PQQKQ RIA RIE RNS AAYXX CITATION NPM RIG 7X8 |
| ID | FETCH-LOGICAL-c368t-7a09128f22fefba8ea7e82d7c2ac93cf4ff2184d4bd9db7df1cc37637b0177c83 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 15 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001128242500002&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2162-237X 2162-2388 |
| IngestDate | Thu Oct 02 06:54:00 EDT 2025 Mon Jul 21 05:59:27 EDT 2025 Sat Nov 29 01:40:26 EST 2025 Tue Nov 18 21:05:09 EST 2025 Wed Aug 27 03:01:49 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | true |
| Issue | 11 |
| 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-c368t-7a09128f22fefba8ea7e82d7c2ac93cf4ff2184d4bd9db7df1cc37637b0177c83 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0001-8661-5695 0000-0002-3625-6599 0000-0002-4499-759X 0000-0001-7624-1861 0000-0002-6639-8547 |
| OpenAccessLink | https://research.tue.nl/en/publications/944c1075-f9f9-4c33-ad21-474d0e791bcf |
| PMID | 37352084 |
| PQID | 2829427713 |
| PQPubID | 23479 |
| PageCount | 13 |
| ParticipantIDs | pubmed_primary_37352084 proquest_miscellaneous_2829427713 crossref_citationtrail_10_1109_TNNLS_2023_3285077 crossref_primary_10_1109_TNNLS_2023_3285077 ieee_primary_10160045 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-11-01 |
| PublicationDateYYYYMMDD | 2024-11-01 |
| PublicationDate_xml | – month: 11 year: 2024 text: 2024-11-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | IEEE transaction on neural networks and learning systems |
| PublicationTitleAbbrev | TNNLS |
| PublicationTitleAlternate | IEEE Trans Neural Netw Learn Syst |
| PublicationYear | 2024 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| References | ref13 Kingma (ref33) ref12 ref36 ref31 Bello (ref18) ref30 ref11 Kim (ref28); 34 ref32 Dai (ref19) ref2 ref1 Kool (ref16) Awad (ref37) ref39 ref38 Joshi (ref20) 2019 Xin (ref21) Nazari (ref14) Vinyals (ref17); 28 Hottung (ref26) Ma (ref15); 34 Hottung (ref24) ref22 (ref34) 2021 Kwon (ref10); 33 ref27 ref29 ref8 ref7 ref9 ref4 ref3 ref6 ref5 Helsgaun (ref35) 2017 Chen (ref23); 32 Lu (ref25) |
| References_xml | – start-page: 1 volume-title: Proc. 24th Eur. Conf. Artif. Intell. ident: ref24 article-title: Neural large neighborhood search for the capacitated vehicle routing problem – ident: ref39 doi: 10.1016/j.ejor.2016.08.012 – ident: ref9 doi: 10.1109/TNNLS.2017.2740224 – ident: ref12 doi: 10.1016/j.cor.2012.07.018 – volume-title: Proc. Int. Conf. Learn. Represent. ident: ref16 article-title: Attention, learn to solve routing problems! – ident: ref2 doi: 10.1007/s10479-009-0657-6 – volume: 34 start-page: 10418 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref28 article-title: Learning collaborative policies to solve NP-hard routing problems – start-page: 374 volume-title: Proc. Int. Conf. Ind. Eng. Oper. Manag. ident: ref37 article-title: An effective genetic algorithm for capacitated vehicle routing problem – ident: ref27 doi: 10.1109/TSMC.2020.2969317 – ident: ref36 doi: 10.1016/j.cor.2021.105643 – ident: ref11 doi: 10.1016/j.ejor.2004.08.018 – start-page: 9861 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref14 article-title: Reinforcement learning for solving the vehicle routing problem – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref25 article-title: A learning-based iterative method for solving vehicle routing problems – ident: ref13 doi: 10.1109/TNNLS.2021.3068828 – ident: ref32 doi: 10.1007/978-3-319-60801-3_27 – ident: ref8 doi: 10.1109/tnnls.2022.3148435 – ident: ref38 doi: 10.1287/ijoc.3.4.376 – start-page: 1 volume-title: Proc. Int. Conf. Mach. Learn. Workshop ident: ref18 article-title: Neural combinatorial optimization with reinforcement learning – volume: 32 start-page: 6281 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref23 article-title: Learning to perform local rewriting for combinatorial optimization – start-page: 6351 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref19 article-title: Learning combinatorial optimization algorithms over graphs – ident: ref29 doi: 10.1109/CVPR.2016.90 – ident: ref22 doi: 10.1109/TITS.2021.3056120 – ident: ref30 doi: 10.5555/3045118.3045167 – year: 2019 ident: ref20 article-title: An efficient graph convolutional network technique for the travelling salesman problem publication-title: arXiv:1906.01227 – year: 2017 ident: ref35 article-title: An extension of the Lin–Kernighan–Helsgaun TSP solver for constrained traveling salesman and vehicle routing problems – ident: ref5 doi: 10.1109/TII.2020.3031409 – ident: ref7 doi: 10.1109/TNNLS.2021.3105905 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref33 article-title: Adam: A method for stochastic optimization – start-page: 12042 volume-title: Proc. AAAI Conf. Artif. Intell. ident: ref21 article-title: Multi-decoder attention model with embedding glimpse for solving vehicle routing problems – ident: ref3 doi: 10.1016/j.eswa.2011.03.088 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref26 article-title: Learning a latent search space for routing problems using variational autoencoders – ident: ref1 doi: 10.1016/0377-2217(92)90192-C – volume-title: Gurobi Optimizer Reference Manual year: 2021 ident: ref34 – volume: 34 start-page: 11096 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref15 article-title: Learning to iteratively solve routing problems with dual-aspect collaborative transformer – volume: 28 start-page: 2692 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref17 article-title: Pointer networks – ident: ref31 doi: 10.1016/S1352-2310(97)00447-0 – ident: ref6 doi: 10.1109/TNNLS.2021.3070584 – ident: ref4 doi: 10.48550/ARXIV.1706.03762 – volume: 33 start-page: 21188 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref10 article-title: POMO: Policy optimization with multiple optima for reinforcement learning |
| SSID | ssj0000605649 |
| Score | 2.5482988 |
| Snippet | While the encoder-decoder structure is widely used in the recent neural construction methods for learning to solve vehicle routing problems (VRPs), they are... |
| SourceID | proquest pubmed crossref ieee |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 15279 |
| SubjectTerms | Costs Decoding Encoder–decoder structure Logic gates neural combinatorial optimization Probability distribution reinforcement learning Routing Search problems Vehicle routing vehicle routing problems (VRPs) |
| Title | Learning Feature Embedding Refiner for Solving Vehicle Routing Problems |
| URI | https://ieeexplore.ieee.org/document/10160045 https://www.ncbi.nlm.nih.gov/pubmed/37352084 https://www.proquest.com/docview/2829427713 |
| Volume | 35 |
| WOSCitedRecordID | wos001128242500002&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: 2162-2388 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000605649 issn: 2162-237X databaseCode: RIE dateStart: 20120101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT8MwDLYAceDC-zFeChI31NEmbZMeEeJxQNPEBtqtahMHkMaG9uD3E6fttAtI3FopSavYiT_HsT-AS4pVJUaYQEY6CeKsDIPS6CKIVGowRefmeprO1yfZ6ajBIOvWyeo-FwYR_eUzbNOjj-WbsZ7TUdm1L4fmMMgqrEopq2StxYFK6IB56uEuj1IecCEHTZJMmF33O52nXpu4wtuCKweCiH1PSAc_fGHTJZvkSVZ-x5ve7txv_fOPt2GzBpjsptKIHVjB0S5sNeQNrF7Le_BQV1Z9Y4QC5xNkd58lGjJl7Bkt5QQyB2hZbzykMwf2iu80IKMrRPTerZhopvvwcn_Xv30MalaFQItUzQJZOIjAleXcoi0LhYVExY3UvNCZ0Da2ltw-E5cmM6U0NtKadiFZusUrtRIHsDYaj_AIWMETTCMbktcTC0zKJBNh5jCd20HJ6LUgauY113XJcWK-GObe9Qiz3IslJ7HktVhacLXo81UV3Piz9T5N-lLLar5bcNHIL3frhYIgxQjH82lOkeOYS-ebt-CwEuyid6MPx7-MegIb7uNxlYp4CmuzyRzPYF1_zz6mk3OnlAN17pXyB7Ih2R4 |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3dT9swED-xMmm8ANsYlK950t6mlMRO4uQRITrQuqhau6pvUWKfB1JpUT_4-_E5SdWXIu0tkWzL8p19v_P57gfwnWJVkRbak4GKvDAtfa_UqvCCJNYYo3VzHU3nqCezLBmP036drO5yYRDRPT7DDn26WL6eqRVdlV25cmgWg7yD3SgMeVCla62vVHwLzWMHeHkQc48LOW7SZPz0aphlvUGH2MI7gicWBhH_npAWgLjSphtWydGsbEeczvJ0D_5zzoewX0NMdl3pxEfYweknOGjoG1i9mz_Dz7q26j9GOHA1R3b7VKImY8b-oKGsQGYhLRvMJnTrwEb4QAMyekRE__2Ki2ZxBH-7t8ObO6_mVfCUiJOlJwsLEnhiODdoyiLBQmLCtVS8UKlQJjSGHD8dljrVpdQmUIrOIVna7StVIr5Aazqb4gmwgkcYB8YnvycUGJVRKvzUojp7hpLZa0PQrGuu6qLjxH0xyZ3z4ae5E0tOYslrsbThx7rPc1Vy483WR7ToGy2r9W7Dt0Z-ud0xFAYppjhbLXKKHYdcWu-8DceVYNe9G3043TLqV_hwN_zdy3v32a8z2LMTCavExHNoLecrvID36mX5uJhfOtV8BQpA230 |
| 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=Learning+Feature+Embedding+Refiner+for+Solving+Vehicle+Routing+Problems&rft.jtitle=IEEE+transaction+on+neural+networks+and+learning+systems&rft.au=Li%2C+Jingwen&rft.au=Ma%2C+Yining&rft.au=Cao%2C+Zhiguang&rft.au=Wu%2C+Yaoxin&rft.date=2024-11-01&rft.issn=2162-2388&rft.eissn=2162-2388&rft.volume=35&rft.issue=11&rft.spage=15279&rft_id=info:doi/10.1109%2FTNNLS.2023.3285077&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2162-237X&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2162-237X&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2162-237X&client=summon |