Automated Algorithm Selection in Single-Objective Continuous Optimization: A Comparative Study of Deep Learning and Landscape Analysis Methods

Raphael Patrick Prager*, Moritz Vinzent Seiler, Heike Trautmann, Pascal Kerschke

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

5 Citations (Scopus)
55 Downloads (Pure)

Abstract

In recent years, feature-based automated algorithm selection using exploratory landscape analysis has demonstrated its great potential in single-objective continuous black-box optimization. However, feature computation is problem-specific and can be costly in terms of computational resources. This paper investigates feature-free approaches that rely on state-of-the-art deep learning techniques operating on either images or point clouds. We show that point-cloud-based strategies, in particular, are highly competitive and also substantially reduce the size of the required solver portfolio. Moreover, we highlight the effect and importance of cost-sensitive learning in automated algorithm selection models.

Original languageEnglish
Title of host publicationParallel Problem Solving from Nature – PPSN XVII
Subtitle of host publication17th International Conference, PPSN 2022, Dortmund, Germany, September 10–14, 2022, Proceedings, Part I
EditorsGünter Rudolph, Anna V. Kononova, Hernán Aguirre, Pascal Kerschke, Gabriela Ochoa, Tea Tušar
Place of PublicationCham
PublisherSpringer
Pages3-17
Number of pages15
ISBN (Electronic)978-3-031-14714-2
ISBN (Print)978-3-031-14713-5
DOIs
Publication statusPublished - 2022
Event17th International Conference on Parallel Problem Solving from Nature, PPSN 2022 - TU Dortmund University, Dortmund, Germany
Duration: 10 Sept 202214 Sept 2022
Conference number: 17

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
Volume13398
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Conference on Parallel Problem Solving from Nature, PPSN 2022
Abbreviated titlePPSN 2022
Country/TerritoryGermany
CityDortmund
Period10/09/2214/09/22

Keywords

  • Automated algorithm selection
  • Continuous optimization
  • Deep learning
  • Exploratory landscape analysis

Fingerprint

Dive into the research topics of 'Automated Algorithm Selection in Single-Objective Continuous Optimization: A Comparative Study of Deep Learning and Landscape Analysis Methods'. Together they form a unique fingerprint.

Cite this