Quantized Zeroth-Order Gradient Tracking Algorithm for Distributed Nonconvex Optimization Under Polyak-Łojasiewicz Condition
This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions. The communication channel between agents is normally constrained by limited bandwidth, and the gradient information is typically unavailable....
Uloženo v:
| Vydáno v: | IEEE transactions on cybernetics Ročník 54; číslo 10; s. 5746 - 5758 |
|---|---|
| Hlavní autoři: | , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
United States
IEEE
01.10.2024
|
| Témata: | |
| ISSN: | 2168-2267, 2168-2275, 2168-2275 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions. The communication channel between agents is normally constrained by limited bandwidth, and the gradient information is typically unavailable. To overcome these limitations, we propose a quantized distributed zeroth-order algorithm, which integrates the deterministic gradient estimator, the standard uniform quantizer, and the distributed gradient tracking algorithm. We establish linear convergence to a global optimal point for the proposed algorithm by assuming Polyak-Łojasiewicz condition for the global cost function and smoothness condition for the local cost functions. Moreover, the proposed algorithm maintains linear convergence at low-data rates with a proper selection of algorithm parameters. Numerical simulations validate the theoretical results. |
|---|---|
| AbstractList | This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions. The communication channel between agents is normally constrained by limited bandwidth, and the gradient information is typically unavailable. To overcome these limitations, we propose a quantized distributed zeroth-order algorithm, which integrates the deterministic gradient estimator, the standard uniform quantizer, and the distributed gradient tracking algorithm. We establish linear convergence to a global optimal point for the proposed algorithm by assuming Polyak-Łojasiewicz condition for the global cost function and smoothness condition for the local cost functions. Moreover, the proposed algorithm maintains linear convergence at low-data rates with a proper selection of algorithm parameters. Numerical simulations validate the theoretical results.This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions. The communication channel between agents is normally constrained by limited bandwidth, and the gradient information is typically unavailable. To overcome these limitations, we propose a quantized distributed zeroth-order algorithm, which integrates the deterministic gradient estimator, the standard uniform quantizer, and the distributed gradient tracking algorithm. We establish linear convergence to a global optimal point for the proposed algorithm by assuming Polyak-Łojasiewicz condition for the global cost function and smoothness condition for the local cost functions. Moreover, the proposed algorithm maintains linear convergence at low-data rates with a proper selection of algorithm parameters. Numerical simulations validate the theoretical results. This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions. The communication channel between agents is normally constrained by limited bandwidth, and the gradient information is typically unavailable. To overcome these limitations, we propose a quantized distributed zeroth-order algorithm, which integrates the deterministic gradient estimator, the standard uniform quantizer, and the distributed gradient tracking algorithm. We establish linear convergence to a global optimal point for the proposed algorithm by assuming Polyak-Łojasiewicz condition for the global cost function and smoothness condition for the local cost functions. Moreover, the proposed algorithm maintains linear convergence at low-data rates with a proper selection of algorithm parameters. Numerical simulations validate the theoretical results. |
| Author | Xu, Lei Shi, Yang Chai, Tianyou Yang, Tao Deng, Chao Yi, Xinlei |
| Author_xml | – sequence: 1 givenname: Lei orcidid: 0000-0003-4753-4774 surname: Xu fullname: Xu, Lei email: 2010345@stu.neu.edu.cn organization: State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang, China – sequence: 2 givenname: Xinlei orcidid: 0000-0003-4299-0471 surname: Yi fullname: Yi, Xinlei email: xinleiyi@mit.edu organization: Laboratory for Information and Decision Systems, Massachusetts Institute of Technology, Cambridge, USA – sequence: 3 givenname: Chao orcidid: 0000-0002-6148-1034 surname: Deng fullname: Deng, Chao email: dengchao_neu@126.com organization: Institute of Advanced Technology, Nanjing University of Posts and Telecommunications, Nanjing, China – sequence: 4 givenname: Yang orcidid: 0000-0003-1337-5322 surname: Shi fullname: Shi, Yang email: yshi@uvic.ca organization: Department of Mechanical Engineering, University of Victoria, Victoria, BC, Canada – sequence: 5 givenname: Tianyou orcidid: 0000-0002-4623-1483 surname: Chai fullname: Chai, Tianyou email: tychai@mail.neu.edu.cn organization: State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang, China – sequence: 6 givenname: Tao orcidid: 0000-0003-4090-8497 surname: Yang fullname: Yang, Tao email: yangtao@mail.neu.edu.cn organization: State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang, China |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/38630570$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9kc1u1DAURi1UREvpAyAhlCWbDP5NnGUZoCBVDEjTBWwsx7lu3Sb2YDtAR2LBw_FeJMwUIRZ4Y8s655Pu_R6iAx88IPSY4AUhuHm-Xn58saCY8gVjkjeU30NHlFSypLQWB3_eVX2ITlK6xtOR01cjH6BDJiuGRY2P0PcPo_bZbaErPkEM-apcxQ5icRZ158DnYh21uXH-sjjtL0N0-WoobIjFS5dydO2YJ_Fd8Cb4L_CtWG2yG9xWZxd8ceHnoPehv9U35c8f4VonB1-d2RbL4Ds3M4_Qfav7BCf7-xhdvH61Xr4pz1dnb5en56VhlObSCGaFAG5rXrW2tWBsI6VmnGsmSGO6umqBiIobLaSmkgDwmgrZWtp2lQB2jJ7tcjcxfB4hZTW4ZKDvtYcwJsUwJ5RNy6wn9OkeHdsBOrWJbtDxVt2tbALqHWBiSCmCVcbl3xPnqF2vCFZzP2ruR839qH0_k0n-Me_C_-c82TkOAP7iBeYVkewXz3yeQA |
| CODEN | ITCEB8 |
| CitedBy_id | crossref_primary_10_1109_TSIPN_2025_3603740 crossref_primary_10_1080_00207179_2025_2537855 crossref_primary_10_1016_j_jfranklin_2025_108014 |
| Cites_doi | 10.1109/TPWRS.2016.2636811 10.1007/s11432-020-3275-3 10.1109/JSAC.2005.843546 10.1109/TII.2012.2219061 10.1109/TCNS.2017.2698261 10.1561/2200000051 10.1109/TAC.2021.3108501 10.1017/CBO9781139020411 10.1109/TAC.2022.3225515 10.1109/TSP.2019.2932876 10.1109/JSTSP.2011.2120593 10.1145/3128572.3140448 10.1137/16M1084316 10.1109/CDC.2008.4738860 10.1109/ALLERTON.2010.5706956 10.1109/TCNS.2020.3024321 10.1109/TAC.2022.3219289 10.1007/978-3-319-46128-1_50 10.1109/TAC.2023.3339439 10.1109/CDC51059.2022.9992989 10.1109/CCDC.2017.7978329 10.1016/j.arcontrol.2019.05.006 10.1137/130943170 10.1007/s11424-021-1231-9 10.1109/TAC.2012.2184199 10.1109/TCNS.2014.2357513 10.1109/TSP.2022.3160238 10.1016/j.automatica.2022.110353 10.1007/978-3-319-91578-4 10.1109/TAC.2020.3031018 10.1109/TAC.2016.2600597 10.1109/CDC.2015.7402509 10.1137/14096668X 10.1109/TAC.2008.2009515 10.1016/j.sysconle.2012.06.004 10.1002/rnc.1147 10.1016/j.acha.2021.12.009 10.1007/s10898-022-01164-w 10.1146/annurev-control-060117-105131 10.1137/19M1258864 10.1109/TSP.2020.2997394 10.1109/PESGM.2018.8586094 |
| ContentType | Journal Article |
| DBID | 97E RIA RIE AAYXX CITATION NPM 7X8 |
| DOI | 10.1109/TCYB.2024.3384924 |
| DatabaseName | IEEE Xplore (IEEE) 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 | Sciences (General) |
| EISSN | 2168-2275 |
| EndPage | 5758 |
| ExternalDocumentID | 38630570 10_1109_TCYB_2024_3384924 10504618 |
| Genre | orig-research Journal Article |
| GrantInformation_xml | – fundername: National Key Research and Development Program of China grantid: 2022YFB3305904 funderid: 10.13039/501100012166 – fundername: National Natural Science Foundation of China grantid: 62133003; 61991403 funderid: 10.13039/501100001809 – fundername: Fundamental Research Funds for the Central Universities grantid: N2324003-05 funderid: 10.13039/501100012226 |
| GroupedDBID | 0R~ 4.4 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK AENEX AGQYO AGSQL AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ EBS EJD HZ~ IFIPE IPLJI JAVBF M43 O9- OCL PQQKQ RIA RIE RNS AAYXX CITATION NPM RIG 7X8 |
| ID | FETCH-LOGICAL-c322t-c53f55e4f746bfbfecf988a344a3519cd76be1564ca58a281ee47258bf2bd65e3 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 6 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001205843000001&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 2168-2275 |
| IngestDate | Sun Nov 09 10:31:20 EST 2025 Mon Jul 21 06:02:24 EDT 2025 Sat Nov 29 02:02:41 EST 2025 Tue Nov 18 21:32:53 EST 2025 Wed Aug 27 03:03:28 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 10 |
| 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-c322t-c53f55e4f746bfbfecf988a344a3519cd76be1564ca58a281ee47258bf2bd65e3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0003-4299-0471 0000-0002-6148-1034 0000-0003-4090-8497 0000-0003-4753-4774 0000-0002-4623-1483 0000-0003-1337-5322 |
| PMID | 38630570 |
| PQID | 3041232027 |
| PQPubID | 23479 |
| PageCount | 13 |
| ParticipantIDs | ieee_primary_10504618 crossref_citationtrail_10_1109_TCYB_2024_3384924 crossref_primary_10_1109_TCYB_2024_3384924 pubmed_primary_38630570 proquest_miscellaneous_3041232027 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-10-01 |
| PublicationDateYYYYMMDD | 2024-10-01 |
| PublicationDate_xml | – month: 10 year: 2024 text: 2024-10-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | IEEE transactions on cybernetics |
| PublicationTitleAbbrev | TCYB |
| PublicationTitleAlternate | IEEE Trans Cybern |
| PublicationYear | 2024 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| References | ref13 ref12 ref34 ref15 ref37 ref14 ref31 ref30 Xu (ref46) 2022 ref11 ref33 ref10 ref32 ref2 ref1 ref17 ref39 ref16 ref38 Fazel (ref35) Li (ref36) ref24 ref45 ref26 ref48 ref25 ref47 ref20 ref42 ref41 ref22 ref44 ref21 ref43 Taheri (ref19) ref28 ref27 ref29 ref8 ref7 Agarwal (ref23) ref9 ref4 ref3 ref6 ref5 ref40 Reisizadeh (ref18); 32 |
| References_xml | – ident: ref4 doi: 10.1109/TPWRS.2016.2636811 – year: 2022 ident: ref46 article-title: Quantized distributed nonconvex optimization algorithms with linear convergence under the Polyak–Łojasiewicz condition publication-title: arXiv:2207.08106 – ident: ref33 doi: 10.1007/s11432-020-3275-3 – ident: ref26 doi: 10.1109/JSAC.2005.843546 – ident: ref5 doi: 10.1109/TII.2012.2219061 – ident: ref15 doi: 10.1109/TCNS.2017.2698261 – ident: ref3 doi: 10.1561/2200000051 – ident: ref37 doi: 10.1109/TAC.2021.3108501 – ident: ref47 doi: 10.1017/CBO9781139020411 – ident: ref48 doi: 10.1109/TAC.2022.3225515 – ident: ref30 doi: 10.1109/TSP.2019.2932876 – ident: ref10 doi: 10.1109/JSTSP.2011.2120593 – ident: ref22 doi: 10.1145/3128572.3140448 – ident: ref14 doi: 10.1137/16M1084316 – ident: ref27 doi: 10.1109/CDC.2008.4738860 – ident: ref11 doi: 10.1109/ALLERTON.2010.5706956 – ident: ref39 doi: 10.1109/TCNS.2020.3024321 – ident: ref21 doi: 10.1109/TAC.2022.3219289 – ident: ref42 doi: 10.1007/978-3-319-46128-1_50 – ident: ref44 doi: 10.1109/TAC.2023.3339439 – ident: ref34 doi: 10.1109/CDC51059.2022.9992989 – volume: 32 start-page: 8388 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref18 article-title: Robust and communication-efficient collaborative learning – ident: ref24 doi: 10.1109/CCDC.2017.7978329 – ident: ref7 doi: 10.1016/j.arcontrol.2019.05.006 – ident: ref9 doi: 10.1137/130943170 – ident: ref32 doi: 10.1007/s11424-021-1231-9 – ident: ref12 doi: 10.1109/TAC.2012.2184199 – ident: ref17 doi: 10.1109/TCNS.2014.2357513 – ident: ref45 doi: 10.1109/TSP.2022.3160238 – ident: ref38 doi: 10.1016/j.automatica.2022.110353 – ident: ref40 doi: 10.1007/978-3-319-91578-4 – ident: ref31 doi: 10.1109/TAC.2020.3031018 – ident: ref29 doi: 10.1109/TAC.2016.2600597 – start-page: 1467 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref35 article-title: Global convergence of policy gradient methods for the linear quadratic regulator – start-page: 28 volume-title: Proc. Annu. Conf. Comput. Learn. Theory ident: ref23 article-title: Optimal algorithms for online convex optimization with multi-point bandit feedback – ident: ref13 doi: 10.1109/CDC.2015.7402509 – ident: ref16 doi: 10.1137/14096668X – ident: ref8 doi: 10.1109/TAC.2008.2009515 – ident: ref28 doi: 10.1016/j.sysconle.2012.06.004 – ident: ref1 doi: 10.1002/rnc.1147 – start-page: 9324 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref19 article-title: Quantized decentralized stochastic learning over directed graphs – ident: ref41 doi: 10.1016/j.acha.2021.12.009 – ident: ref43 doi: 10.1007/s10898-022-01164-w – ident: ref6 doi: 10.1146/annurev-control-060117-105131 – ident: ref20 doi: 10.1137/19M1258864 – start-page: 5569 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref36 article-title: A simple proximal stochastic gradient method for nonsmooth nonconvex optimization – ident: ref25 doi: 10.1109/TSP.2020.2997394 – ident: ref2 doi: 10.1109/PESGM.2018.8586094 |
| SSID | ssj0000816898 |
| Score | 2.4051027 |
| Snippet | This article focuses on distributed nonconvex optimization by exchanging information between agents to minimize the average of local nonconvex cost functions.... |
| SourceID | proquest pubmed crossref ieee |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 5746 |
| SubjectTerms | Convergence Cost function Deep learning Distributed algorithms Gradient methods Gradient tracking algorithm linear convergence nonconvex optimization Quantization (signal) uniform quantizer zeroth-order algorithm |
| Title | Quantized Zeroth-Order Gradient Tracking Algorithm for Distributed Nonconvex Optimization Under Polyak-Łojasiewicz Condition |
| URI | https://ieeexplore.ieee.org/document/10504618 https://www.ncbi.nlm.nih.gov/pubmed/38630570 https://www.proquest.com/docview/3041232027 |
| Volume | 54 |
| WOSCitedRecordID | wos001205843000001&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/eLvHCXMwlV1Lb9QwELZohVAvhUKhy6MyEgdASskmduwcy0LhgJYiFbRwiWxnDIFtgrJZHpU48OP4X8w42VUvReKWg51EmZnMjGfm-xh7kMfe6MT7qLSpiESifJRnpYusAAXe50YGtM93r9R0qmez_HgYVg-zMAAQms_ggC5DLb9s3JKOytDCJeGD6w22oVTWD2utD1QCg0Tgvk3wIsKwQg1VzHGcPzmZvH-K2WAiDjAnE5h0bLErqc5Q24mm-JxLChwrF4ebwe0cXf3PF77Gtof4kh_2CrHDLkF9ne0MFrzgDweY6Uc32K83S_yq1RmU_AO0DR1xEgwnf9GGLrCOoxtzdJDOD-cfm7bqPp1yjHD5M4LaJZYs3Dht6tC2_oO_xl_P6TDTyQOZEj9u5j_Nl-jP7-azWVTwvXJnfNJQiRzX7LK3R89PJi-jgYwhcmjzXeRk6qUE4ZXIrLcenM-1NqkQhjj-XKkyC4Q844zUJtFjAKESqa1PbJlJSG-yzbqpYY9x7WVmMbHMMNkTNhFGazuWKLu4RDdgkxGLV_Io3IBUToQZ8yJkLHFekDQLkmYxSHPEHq-3fO1hOv61eJdEdW5hL6URu7-SeoFGRpUTU0OzXBQpoZIR07wasVu9Oqx3r7To9gV3vcO26OF9A-Bdttm1S7jHLrtvXbVo91GTZ3o_aPJf75LvjA |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Lb9QwEB5BQbQXoFBgeRqJAyClZB07cY5loRSxLEVaUOES2Y5dAtsEZbM8KnHgx_G_8DjeVS9F4paDHVmZmczDM98H8CCPrRTU2qhUCYsYzWyUp6WOFDOZsTaX3KN9vh9nk4k4OMj3w7C6n4UxxvjmM7ONj_4uv2z0AktlzsI54oOLs3COM0bjflxrVVLxHBKe_Za6h8gFFlm4xxzG-ZPp6MNTlw9Stu2yMubSjg24kIjU6TsSFZ9wSp5l5fSA0zue3Uv_eeTLcDFEmGSnV4lNOGPqK7AZbHhOHgag6UdX4dfbhfuu1bEpyUfTNljkRCBO8qL1fWAdcY5MYymd7MwOm7bqPh0RF-OSZwi2izxZbuOkqX3j-g_yxv18jsJUJ_F0SmS_mf2UX6I_v5vPcl6Z75U-JqMGL8ndmi14t_t8OtqLAh1DpJ3Vd5HmieXcMJuxVFlljba5EDJhTCLLny6zVBnEntGSC0nF0BiWUS6UpapMuUmuwVrd1OYGEGF5qlxqmbp0jynKpBBqyJ3s4tI5AkUHEC_lUeiAVY6UGbPC5yxxXqA0C5RmEaQ5gMerLV97oI5_Ld5CUZ1Y2EtpAPeXUi-cmeHdiaxNs5gXCeKSIdd8NoDrvTqsdi-16OYpb70H63vT1-Ni_HLy6hZs4EH6dsDbsNa1C3MHzutvXTVv73p9_gvjUvHr |
| 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=Quantized+Zeroth-Order+Gradient+Tracking+Algorithm+for+Distributed+Nonconvex+Optimization+Under+Polyak-%C5%81ojasiewicz+Condition&rft.jtitle=IEEE+transactions+on+cybernetics&rft.au=Xu%2C+Lei&rft.au=Yi%2C+Xinlei&rft.au=Deng%2C+Chao&rft.au=Shi%2C+Yang&rft.date=2024-10-01&rft.pub=IEEE&rft.issn=2168-2267&rft.volume=54&rft.issue=10&rft.spage=5746&rft.epage=5758&rft_id=info:doi/10.1109%2FTCYB.2024.3384924&rft_id=info%3Apmid%2F38630570&rft.externalDocID=10504618 |
| 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 |