Algorithmic construction of optimal designs on compact sets for concave and differentiable criteria

Abstract : We consider the problem of construction of optimal experimental designs (approximate theory) on a compact subset X of Rd with nonempty interior, for a concave and Lipschitz diff erentiable design criterion Phi (.) based on the information matrix. The proposed algorithm combines (a) convex optimization for the determination of optimal weights on a support set, (b) sequential updating of this support using local optimization, and (c) finding new support candidates using properties of the directional derivative of Phi(.). The algorithm makes use of the compactness of X and relies on a fi nite grid Xl C X for checking optimality. By exploiting the Lipschitz continuity of the directional derivatives of Phi(.), effi ciency bounds on X are obtained and epsilon-optimality on X is guaranteed. The eff ectiveness of the method is illustrated on a series of examples.
Type de document :
Article dans une revue
Journal of Statistical Planning and Inference, Elsevier, 2014, 154, pp.141-155. 〈10.1016/j.jspi.2014.04.005〉
Liste complète des métadonnées

Littérature citée [52 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01001706
Contributeur : Luc Pronzato <>
Soumis le : mercredi 4 juin 2014 - 17:09:03
Dernière modification le : mercredi 7 septembre 2016 - 10:50:05
Document(s) archivé(s) le : jeudi 4 septembre 2014 - 13:05:35

Fichier

Opt_design-V5.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Luc Pronzato, Anatoly Zhigljavsky. Algorithmic construction of optimal designs on compact sets for concave and differentiable criteria. Journal of Statistical Planning and Inference, Elsevier, 2014, 154, pp.141-155. 〈10.1016/j.jspi.2014.04.005〉. 〈hal-01001706〉

Partager

Métriques

Consultations de
la notice

252

Téléchargements du document

156