Landscape analysis-based automated algorithm selection

Simple item page
dc.contributor.advisorEngelbrecht, Andriesen_ZA
dc.contributor.authorLang, Ryan Dieteren_ZA
dc.contributor.otherStellenbosch University. Faculty of Science. Dept. of Computer Science.en_ZA
dc.date.accessioned2024-02-24T12:58:06Z
dc.date.accessioned2024-04-26T10:30:58Z
dc.date.available2024-02-24T12:58:06Z
dc.date.available2024-04-26T10:30:58Z
dc.date.issued2024-03
dc.descriptionThesis (MSc)--Stellenbosch University, 2024.en_ZA
dc.description.abstractENGLISH ABSTRACT: The algorithm selection problem, which was developed by Rice, refers to the chal- lenge of choosing the best algorithm available to solve a specific p roblem. The framework put forth by Rice includes four spaces: the problem space, the algorithm space, the characteristic space, and the performance space. In this thesis, the focus is on the problem space of continuous-valued single-objective, boundary-constrained optimisation problems. Landscape analysis, which is a set of techniques that utilises mathematical and statistical methods to characterise the properties of optimisation problems, can be used to describe the characteristic space. Selection of the most effective algorithm to solve optimisation problems is a complex task, because meta- heuristics have varying strengths. Therefore, a data-driven approach that utilises landscape analysis techniques is employed in this thesis to create an automated algorithm selector. This thesis proposes enhancements to the characteristic space by critically evaluat- ing the reliability of the methods used in generating landscape analysis measures. Additionally, a new benchmark suite, which is more representative of the problem space than commonly used benchmark suites in the literature, is proposed. An investigation into the impact on the performance of hybrid metaheuristics created by combining the sampling algorithms used to calculate landscape analysis features with standard metaheuristics, which are used to solve the optimisation problem, is conducted. By combining the proposed improvements to the characteristic, algo- rithm, and problem spaces of the algorithm selection framework, this thesis concludes with a landscape analysis-based automated algorithm selection model that outper- forms the current automated algorithm selection models found in the literature in terms of performance and generalisability. Furthermore, this thesis explores the use of automated machine learning and hybrid metaheuristics in automated algorithm selection.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: Die algoritme seleksie probleem, ontwikkel deur Rice, verwys na die uitdaging om die beste beskikbare algoritme te selekteer om ’n spesifieke p robleem op te l os. Die raamwerk wat deur Rice voorgestel is, sluit vier ruimtes in: die probleem ruimte, die algoritme ruimte, die eienskap ruimte, en die prestasie ruimte. Die fokus van hierdie tesis is op die probleem ruimtes van kontinue-waarde, enkel-doelwit, grens-beperkde optimeringsprobleme. Landskap analise, ’n versameling van tegnieke wat wiskun- dige en statistiese metodes aanwend om die eienskappe van optimeringsprobleme te karakteriseer, kan gebruik word om die eienskap ruimte te beskryf. Seleksie van die mees effektiewe a lgoritme v ir o ptimeringsprobleme i s ’ n k omplekse t aak, aan- gesien meta-heuristieke verskillende sterkpunte het. Daarom wend hierdie tesis ’n data-gedrewe benadering aan wat landskap analise tegniekie gebruik om ’n geouto- matiseerde algoritme keurder te ontwikkel. Hierdie tesis stel verbeteringe voor vir die eienskap ruimte deur kritiese evaluasie van die betroubaarheid van metodes wat gebruik word om landskap analise maat- stawe te genereer. Addisioneel word ’n nuwe probleem suite voorgestel wat meer verteenwoordigend van die probleem ruimte is as bestaande probleem suites wat in die literatuur beskikbaar is. ’n Ondersoek word ingestel na die prestasie van hibriede meta-heuristieke wat gevorm word deur monsternemingalgoritmes wat gebruik word om landskap eienskappe te bereken met standaard meta-heuristieke te kombineer. Deur die voorgestelde verbeteringe in die eienskap, algoritme, en probleem ruimtes van die algoritme seleksie raamwerk in te sluit, sluit hierdie tesis af met ’n landskap analise gebaseerde outomatiese algoritme seleksie model wat beter presteer in terme van akkuraathied en veralgemening as huidige outomatiese algoritme seleksie mo- delle wat in die literatuur gevind kan word. Verder verken die tesis die gebruik van geoutomatiseerde masjienleer en hibriede meta-heuristieke in outomatiese algoritme seleksie.af_ZA
dc.description.versionMastersen_ZA
dc.format.extentxviii, 157 pages : illustrations (some color)en_ZA
dc.identifier.urihttps://scholar.sun.ac.za/handle/10019.1/130243
dc.language.isoen_ZAen_ZA
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subject.lcshComputer algorithms -- Mathematical modelsen_ZA
dc.subject.lcshAlgorithm selection problem -- Measurementen_ZA
dc.subject.lcshLandscapes -- Analysis -- Computer programsen_ZA
dc.subject.lcshMetaheuristicsen_ZA
dc.subject.nameUCTDen_ZA
dc.titleLandscape analysis-based automated algorithm selectionen_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
lang_landscape_2024.pdf
Size:
17.47 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Masters Degrees (Computer Science)