2026-07-23T20:02:51Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:20723 2024-08-29T08:52:53Z ddc:004 open_access

Frugal Algorithm Selection (Short Paper) Kuş, Erdem Akgün, Özgür Dang, Nguyen Miguel, Ian Algorithm Selection Active Learning When solving decision and optimisation problems, many competing algorithms (model and solver choices) have complementary strengths. Typically, there is no single algorithm that works well for all instances of a problem. Automated algorithm selection has been shown to work very well for choosing a suitable algorithm for a given instance. However, the cost of training can be prohibitively large due to running candidate algorithms on a representative set of training instances. In this work, we explore reducing this cost by choosing a subset of the training instances on which to train. We approach this problem in three ways: using active learning to decide based on prediction uncertainty, augmenting the algorithm predictors with a timeout predictor, and collecting training data using a progressively increasing timeout. We evaluate combinations of these approaches on six datasets from ASLib and present the reduction in labelling cost achieved by each option. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Erdem Kuş and Özgür Akgün and Nguyen Dang and Ian Miguel 2024 Is Part Of LIPIcs, Volume 307, 30th International Conference on Principles and Practice of Constraint Programming (CP 2024) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.CP.2024.38 urn:nbn:de:0030-drops-207239 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.CP.2024.38 eng https://creativecommons.org/licenses/by/4.0/legalcode