Preconditioning of a Generalized Forward-Backward Splitting and Application to Optimization on Graphs - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue SIAM Journal on Imaging Sciences Année : 2015

Preconditioning of a Generalized Forward-Backward Splitting and Application to Optimization on Graphs

Préconditionnement d'une généralisation de l'algorithme explicite-implicite, et application à l'optimisation sur des graphes

Résumé

We present a preconditioning of a generalized forward-backward splitting algorithm for nding a zero of a sum of maximally monotone operators \sum_{i=1}^n A_i + B with B cocoercive, involving only the computation of B and of the resolvent of each A i separately. is allows in particular to minimize functionals of the form \sum_{i=1}^n g_i + f with f smooth. By adapting the underlying metric, such preconditioning can serve two practical purposes: rst, it might accelerate the convergence, or second, it might simplify the computation of the resolvent of A i for some i. In addition, in many cases of interest, our preconditioning strategy allows the economy of storage and computation concerning some auxiliary variables. In particular, we show how this approach can handle large-scale, nonsmooth, convex optimization problems structured on graphs, which arises in many image processing or learning applications, and that it compares favorably to alternatives in the literature.
Fichier principal
Vignette du fichier
PGFB-preprint.pdf (5.03 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-01144566 , version 1 (22-04-2015)
hal-01144566 , version 2 (14-05-2015)
hal-01144566 , version 3 (06-07-2015)

Identifiants

Citer

Hugo Raguet Raguet, Loic Landrieu. Preconditioning of a Generalized Forward-Backward Splitting and Application to Optimization on Graphs. SIAM Journal on Imaging Sciences, 2015, 8 (4), pp.2706-2739. ⟨10.1137/15M1018253⟩. ⟨hal-01144566v3⟩
489 Consultations
398 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More