Trajectory-based Algorithm Selection with Warm-starting
Landscape-aware algorithm selection approaches have so far mostly been relying on landscape feature extraction as a preprocessing step, independent of the execution of optimization algorithms in the portfolio. This introduces a significant overhead in computational cost for many practical applicatio...
Uloženo v:
| Vydáno v: | 2022 IEEE Congress on Evolutionary Computation (CEC) s. 1 - 8 |
|---|---|
| Hlavní autoři: | , , , , , |
| Médium: | Konferenční příspěvek |
| Jazyk: | angličtina |
| Vydáno: |
IEEE
18.07.2022
|
| Témata: | |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | Landscape-aware algorithm selection approaches have so far mostly been relying on landscape feature extraction as a preprocessing step, independent of the execution of optimization algorithms in the portfolio. This introduces a significant overhead in computational cost for many practical applications, as features are extracted and computed via sampling and evaluating the problem instance at hand, similarly to what the optimization algorithm would perform anyway within its search trajectory. As suggested in [Jankovic et al., EvoAPP 2021], trajectory-based algorithm selection circumvents the problem of costly feature extraction by computing landscape features from points that a solver sampled and evaluated during the optimization process. Features computed in this manner are used to train algorithm performance regression models, upon which a per-run algorithm selector is then built. In this work, we apply the trajectory-based approach onto a portfolio of five algorithms. We study the quality and accuracy of performance regression and algorithm selection models in the scenario of predicting different algorithm performances after a fixed budget of function evaluations. We rely on landscape features of the problem instance computed using one portion of the aforementioned budget of the same function evaluations. Moreover, we consider the possibility of switching between the solvers once, which requires them to be warm-started, i.e. when we switch, the second solver continues the optimization process already being initialized appropriately by making use of the information collected by the first solver. In this new context, we show promising performance of the trajectory-based per-run algorithm selection with warm-starting. |
|---|---|
| AbstractList | Landscape-aware algorithm selection approaches have so far mostly been relying on landscape feature extraction as a preprocessing step, independent of the execution of optimization algorithms in the portfolio. This introduces a significant overhead in computational cost for many practical applications, as features are extracted and computed via sampling and evaluating the problem instance at hand, similarly to what the optimization algorithm would perform anyway within its search trajectory. As suggested in [Jankovic et al., EvoAPP 2021], trajectory-based algorithm selection circumvents the problem of costly feature extraction by computing landscape features from points that a solver sampled and evaluated during the optimization process. Features computed in this manner are used to train algorithm performance regression models, upon which a per-run algorithm selector is then built. In this work, we apply the trajectory-based approach onto a portfolio of five algorithms. We study the quality and accuracy of performance regression and algorithm selection models in the scenario of predicting different algorithm performances after a fixed budget of function evaluations. We rely on landscape features of the problem instance computed using one portion of the aforementioned budget of the same function evaluations. Moreover, we consider the possibility of switching between the solvers once, which requires them to be warm-started, i.e. when we switch, the second solver continues the optimization process already being initialized appropriately by making use of the information collected by the first solver. In this new context, we show promising performance of the trajectory-based per-run algorithm selection with warm-starting. |
| Author | de Nobel, Jacob Jankovic, Anja Vermetten, Diederick Eftimov, Tome Kostovska, Ana Doerr, Carola |
| Author_xml | – sequence: 1 givenname: Anja surname: Jankovic fullname: Jankovic, Anja email: Anja.Jankovic@lip6.fr organization: CNRS, LIP6, Sorbonne Université,Paris,France – sequence: 2 givenname: Diederick surname: Vermetten fullname: Vermetten, Diederick email: d.l.vermetten@liacs.leidenuniv.nl organization: Leiden University,Netherlands – sequence: 3 givenname: Ana surname: Kostovska fullname: Kostovska, Ana email: ana.kostovska@ijs.si organization: Jožef Stefan Institute, Jožef Stefan International Postgraduate School,Department of Knowledge Technologies,Ljubljana,Slovenia – sequence: 4 givenname: Jacob surname: de Nobel fullname: de Nobel, Jacob email: j.p.de.nobel@liacs.leidenuniv.nl organization: Leiden University,Netherlands – sequence: 5 givenname: Tome surname: Eftimov fullname: Eftimov, Tome email: tome.eftimov@ijs.si organization: Jožef Stefan Institute,Computer Systems Department,Ljubljana,Slovenia – sequence: 6 givenname: Carola surname: Doerr fullname: Doerr, Carola email: Carola.Doerr@lip6.fr organization: CNRS, LIP6, Sorbonne Université,Paris,France |
| BookMark | eNotj9tKw0AURUdQUGu-QIT8wMQz95nHEuoFCj5Y8bGcJCdtSi4yGZD-vQH7tNhs1oZ9z67HaSTGngQUQkB4LjelMWBNIUHKIni3QF6xLDgvrDXaOvDulmXzfAIA6SVoZe-Y20U8UZ2meOYVztTk6_4wxS4dh_yT-qXppjH_XXL-jXHgc8KYuvHwwG5a7GfKLlyxr5fNrnzj24_X93K95UfpdeJS2kYHpVFXAS2B91C3RKRNS0KBtIhYW3KNp9AKqiuqgjEavUVFWrRqxR7_d7vF2v_EbsB43l_uqT9mPUib |
| ContentType | Conference Proceeding |
| DBID | 6IE 6IL CBEJK RIE RIL |
| DOI | 10.1109/CEC55065.2022.9870222 |
| DatabaseName | IEEE Electronic Library (IEL) Conference Proceedings IEEE Xplore POP ALL IEEE Xplore All Conference Proceedings IEEE/IET Electronic Library (IEL) (UW System Shared) IEEE Proceedings Order Plans (POP All) 1998-Present |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE/IET Electronic Library (IEL) (UW System Shared) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| EISBN | 9781665467087 1665467088 |
| EndPage | 8 |
| ExternalDocumentID | 9870222 |
| Genre | orig-research |
| GrantInformation_xml | – fundername: EC grantid: 952215 (TAILOR) funderid: 10.13039/501100000780 – fundername: Slovenian Research Agency grantid: P2-0098,P2-0103,PR-09773 funderid: 10.13039/501100004329 |
| GroupedDBID | 6IE 6IF 6IL 6IN AAWTH ABLEC ALMA_UNASSIGNED_HOLDINGS BEFXN BFFAM BGNUA BKEBE BPEOZ CBEJK IEGSK OCL RIE RIL |
| ID | FETCH-LOGICAL-h284t-226d4934a4b9a6e0880cfeee45fe13026aaac6e7d8e9f1ecbeb9554a86a3e41f3 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 6 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000859282000019&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| IngestDate | Wed Aug 27 02:19:12 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | false |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-h284t-226d4934a4b9a6e0880cfeee45fe13026aaac6e7d8e9f1ecbeb9554a86a3e41f3 |
| OpenAccessLink | https://hal.sorbonne-universite.fr/hal-03774161 |
| PageCount | 8 |
| ParticipantIDs | ieee_primary_9870222 |
| PublicationCentury | 2000 |
| PublicationDate | 2022-July-18 |
| PublicationDateYYYYMMDD | 2022-07-18 |
| PublicationDate_xml | – month: 07 year: 2022 text: 2022-July-18 day: 18 |
| PublicationDecade | 2020 |
| PublicationTitle | 2022 IEEE Congress on Evolutionary Computation (CEC) |
| PublicationTitleAbbrev | CEC |
| PublicationYear | 2022 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| SSID | ssj0002820436 |
| Score | 1.9226768 |
| Snippet | Landscape-aware algorithm selection approaches have so far mostly been relying on landscape feature extraction as a preprocessing step, independent of the... |
| SourceID | ieee |
| SourceType | Publisher |
| StartPage | 1 |
| SubjectTerms | black-box optimization Computational modeling dynamic algorithm selection evolutionary computation exploratory landscape analysis Feature extraction Heuristic algorithms Prediction algorithms Predictive models Search problems Switches |
| Title | Trajectory-based Algorithm Selection with Warm-starting |
| URI | https://ieeexplore.ieee.org/document/9870222 |
| WOSCitedRecordID | wos000859282000019&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 | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NSwMxEB3a4sGTSit-swePpt3vbI5SWjxIKajYW5lNJq1iW6lbwX_vJF0qghdvYSHsZgJ581523gBcY2ZQRliKuNQJExQsBIbGCiYTNsxKxdzHW-bfy9GomEzUuAE3u1oYIvI_n1HXDf1dvlnpjZPKesyPHT9pQlPKfFurtdNTmDo4N_W6SCcKVa8_6HP6nWdMAuO4W8_91UTFY8jw4H9vP4TOTzFeMN7BzBE0aNkGyRjz6gX3L-GAyAS3b7MVE_35InjwrW043oETWYNnXC8E54DOLmDWgafh4LF_J-oeCGLOwFEJzo5MqpIU01JhTnwmhNryp6WZJXfnmCOizkmagpSNSJfE8c1SLHJMKI1scgyt5WpJJxDEiVXk-8QzLstModGG0wFnuE6xMvkptN2ip-9bm4tpvd6zvx-fw76Lq5M5o-ICWtV6Q5ewpz-rl4_1ld-bb166j-g |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NSwMxEB1qFfSk0orf7sGjafcj2d0cpbRUrKVgxd5KdjNpFduV2gr-eyfpUhG8eAsLYTcTyJv3svMG4FoJrZJAZSzM8ogIikqZ8rVhRCaMLzJJ3MdZ5veSfj8djeSgAjebWhhEdD-fYcMO3V2-LvKVlcqaxI8tP9mCbcF56K-rtTaKCpEH66delukEvmy22i1KwGNBNDAMG-XsX21UHIp09v_3_gOo_5TjeYMN0BxCBec1SAhlXp3k_sUsFGnv9m1SENWfzrxH19yGIu5ZmdV7VosZoyzQGgZM6vDUaQ9bXVZ2QWBTgo4lo_xIcxlxxTOpYqRTwc8NfRoXBu2tY6yUymNMdIrSBJhnSBEWXKWxipAHJjqC6ryY4zF4YWQkuk7xhMyJkErnmhICa7mOodTxCdTsosfva6OLcbne078fX8Fud_jQG_fu-vdnsGdjbEXPID2H6nKxwgvYyT-XLx-LS7dP3w28ky8 |
| 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%3Abook&rft.genre=proceeding&rft.title=2022+IEEE+Congress+on+Evolutionary+Computation+%28CEC%29&rft.atitle=Trajectory-based+Algorithm+Selection+with+Warm-starting&rft.au=Jankovic%2C+Anja&rft.au=Vermetten%2C+Diederick&rft.au=Kostovska%2C+Ana&rft.au=de+Nobel%2C+Jacob&rft.date=2022-07-18&rft.pub=IEEE&rft.spage=1&rft.epage=8&rft_id=info:doi/10.1109%2FCEC55065.2022.9870222&rft.externalDocID=9870222 |