Landscape features and automated algorithm selection for multi-objective interpolated continuous optimisation problems - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Landscape features and automated algorithm selection for multi-objective interpolated continuous optimisation problems

Benjamin Lacroix
  • Fonction : Auteur
  • PersonId : 1108178
Alexandru-Ciprian Zăvoianu
  • Fonction : Auteur
  • PersonId : 1108179
John Mccall
  • Fonction : Auteur
  • PersonId : 1108180

Résumé

In this paper, we demonstrate the application of features from landscape analysis, initially proposed for multi-objective combinatorial optimisation, to a benchmark set of 1200 randomly-generated multiobjective interpolated continuous optimisation problems (MO-ICOPs).We also explore the benefits of evaluating the considered landscape features on the basis of a fixed-size sampling of the search space. This allows fine control over cost when aiming for an efficient application of feature-based automated performance prediction and algorithm selection. While previous work shows that the parameters used to generate MO-ICOPs are able to discriminate the convergence behaviour of four state-of-the-art multi-objective evolutionary algorithms, our experiments reveal that the proposed (black-box) landscape features used as predictors deliver a similar accuracy when combined with a classification model. In addition, we analyse the relative importance of each feature for performance prediction and algorithm selection.
Fichier principal
Vignette du fichier
gecco21.pdf (725.94 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03325676 , version 1 (25-08-2021)

Identifiants

Citer

Arnaud Liefooghe, Sébastien Verel, Benjamin Lacroix, Alexandru-Ciprian Zăvoianu, John Mccall. Landscape features and automated algorithm selection for multi-objective interpolated continuous optimisation problems. GECCO 2021 - The Genetic and Evolutionary Computation Conference, Jul 2021, Lille / Virtual, France. pp.421-429, ⟨10.1145/3449639.3459353⟩. ⟨hal-03325676⟩
69 Consultations
80 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More