hal-00721886, version 1
Convergence and efficiency of the Wang-Landau algorithm
(30/07/2012)
Résumé : We analyze the convergence properties of the Wang-Landau algorithm. This sampling method belongs to the general class of adaptive importance sampling strategies which use the free energy along a chosen reaction coordinate as a bias. Such algorithms are very helpful to enhance the sampling properties of Markov Chain Monte Carlo algorithms, when the dynamic is metastable. We prove that the Wang-Landau algorithm converges with an associated central limit theorem, and we provide an analysis of the efficiency of the algorithm in a metastable situation.
- 1 :
- Télécom ParisTech – CNRS : UMR5141
- 2 :
- Ecole des Ponts ParisTech
- 3 :
- Institut national de la recherche agronomique (INRA) : UMR0518 – AgroParisTech
- 4 :
- Ecole des Ponts ParisTech – INRIA
- Domaine : Mathématiques/Probabilités
Statistiques/Théorie - Commentaire : This work is supported by the French National Research Agency under the grants ANR-09-BLAN-0216-01 (MEGAS) and ANR-08-BLAN-0218 (BigMC)
- hal-00721886, version 1
- http://hal.inria.fr/hal-00721886
- oai:hal.inria.fr:hal-00721886
- Contributeur :
- Soumis le : Mardi 31 Juillet 2012, 08:19:43
- Dernière modification le : Mercredi 9 Janvier 2013, 10:18:34



Documents associés
Exporter