Quantile-like measures on multi-dimensional distributions of closed sets - Archive ouverte HAL Accéder directement au contenu
Cours Année : 2023

Quantile-like measures on multi-dimensional distributions of closed sets

Johann Dreo

Résumé

The Empirical Attainment Function is a 2D empirical distribution function, which is actually a distribution of convergence trajectories. It can be efficiently computed, using the monotonic nature of those trajectories. We argue that it extends both the classical fixed-target and fixed-budget ECDF, which are the current gold-standard to assess black-box heuristic search performances. We argue that, in addition to capturing more information, if is also less prone to approximation error, as it does not need a pre-defined target set.
Fichier principal
Vignette du fichier
2021-06-10_Dreo__EAF.pdf (4.72 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04110682 , version 1 (30-05-2023)

Identifiants

  • HAL Id : hal-04110682 , version 1

Citer

Johann Dreo. Quantile-like measures on multi-dimensional distributions of closed sets: Application in stochastic optimization. Doctoral. Innovation, Design, and Engineering, Västerås [Sweden], France. 2023. ⟨hal-04110682⟩

Collections

PASTEUR TDS-MACS
11 Consultations
3 Téléchargements

Partager

Gmail Facebook X LinkedIn More