Decomposition-based multi-objective landscape features and automated algorithm selection - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Decomposition-based multi-objective landscape features and automated algorithm selection

Abstract

Landscape analysis is of fundamental interest for improving our understanding on the behavior of evolutionary search, and for developing general-purpose automated solvers based on techniques from statistics and machine learning. In this paper, we push a step towards the development of a landscape-aware approach by proposing a set of landscape features for multi-objective combinatorial optimization, by decomposing the original multi-objective problem into a set of single-objective sub-problems. Based on a comprehensive set of bi-objective Open image in new window and three variants of the state-of-the-art MOEA/D algorithm, we study the association between the proposed features, the global properties of the considered landscapes, and algorithm performance. We also show that decomposition-based features can be integrated into an automated approach for predicting algorithm performance and selecting the most accurate one on blind instances. In particular, our study reveals that such a landscape-aware approach is substantially better than the single best solver computed over the three considered MOEA/D variants.
Fichier principal
Vignette du fichier
cosson_evocop2021.pdf (3.61 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03331977 , version 1 (02-03-2023)

Identifiers

Cite

Raphaël Cosson, Bilel Derbel, Arnaud Liefooghe, Hernán Aguirre, Kiyoshi Tanaka, et al.. Decomposition-based multi-objective landscape features and automated algorithm selection. EvoCOP 2021 - 21st European Conference on Evolutionary Computation in Combinatorial Optimization, 2021, Virtual Event, Spain. pp.34-50, ⟨10.1007/978-3-030-72904-2_3⟩. ⟨hal-03331977⟩
99 View
24 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More