A4 Article in conference proceedings
Feature-Based Benchmarking of Distance-Based Multi/Many-objective Optimisation Problems : A Machine Learning Perspective (2023)


Liefooghe, A., Verel, S., Chugh, T., Fieldsend, J., Allmendinger, R., & Miettinen, K. (2023). Feature-Based Benchmarking of Distance-Based Multi/Many-objective Optimisation Problems : A Machine Learning Perspective. In M. Emmerich, A. Deutz, H. Wang, A. V. Kononova, B. Naujoks, K. Li, K. Miettinen, & I. Yevseyeva (Eds.), Evolutionary Multi-Criterion Optimization : 12th International Conference, EMO 2023, Leiden, The Netherlands, March 20–24, 2023, Proceedings (pp. 260-273). Springer. Lecture Notes in Computer Science, 13970. https://doi.org/10.1007/978-3-031-27250-9_19


JYU authors or editors


Publication details

All authors or editors: Liefooghe, Arnaud; Verel, Sébastien; Chugh, Tinkle; Fieldsend, Jonathan; Allmendinger, Richard; Miettinen, Kaisa

Parent publication: Evolutionary Multi-Criterion Optimization : 12th International Conference, EMO 2023, Leiden, The Netherlands, March 20–24, 2023, Proceedings

Parent publication editors: Emmerich, Michael; Deutz, André; Wang, Hao; Kononova, Anna V.; Naujoks, Boris; Li, Ke; Miettinen, Kaisa; Yevseyeva, Iryna

Place and date of conference: Leiden, The Netherlands, 20.-24.3.2023

ISBN: 978-3-031-27249-3

eISBN: 978-3-031-27250-9

Journal or series: Lecture Notes in Computer Science

ISSN: 0302-9743

eISSN: 1611-3349

Publication year: 2023

Publication date: 21/02/2023

Number in series: 13970

Pages range: 260-273

Number of pages in the book: 636

Publisher: Springer

Place of Publication: Cham

Publication country: Switzerland

Publication language: English

DOI: https://doi.org/10.1007/978-3-031-27250-9_19

Publication open access: Not open

Publication channel open access: 

Publication is parallel published (JYX): https://jyx.jyu.fi/handle/123456789/87040


Abstract

We consider the application of machine learning techniques to gain insights into the effect of problem features on algorithm performance, and to automate the task of algorithm selection for distance-based multi- and many-objective optimisation problems. This is the most extensive benchmark study of such problems to date. The problem features can be set directly by the problem generator, and include e.g. the number of variables, objectives, local fronts, and disconnected Pareto sets. Using 945 problem configurations (leading to 28350
instances) of varying complexity, we find that the problem features and the available optimisation budget (i) affect the considered algorithms (NSGA-II, IBEA, MOEA/D, and random search) in different ways and that (ii) it is possible to recommend a relevant algorithm based on problem features.


Keywords: multi-objective optimisation; algorithms; benchmarking; machine learning

Free keywords: multi/many-objective distance problems; feature-based performance prediction; automated algorithm selection


Contributing organizations


Ministry reporting: Yes

Reporting Year: 2023

Preliminary JUFO rating: 1


Last updated on 2024-22-04 at 14:44