We propose a new gradient projection algorithm that compares favorably with the fastest algorithms available to date for ℓ1-constrained sparse recovery from noisy data, both in the compressed sensing and inverse problem frameworks. The method exploits a line-search along the feasible direction and an adaptive steplength selection based on recent strategies for the alternation of the well-known Barzilai–Borwein rules. The convergence of the proposed approach is discussed and a computational study on both well conditioned and ill-conditioned problems is carried out for performance evaluations in comparison with five other algorithms proposed in the literature.
Accelerating gradient projection methods for $ell_1$-constrained signal recovery by steplength selection rules / Loris, I; Bertero, M; DE MOL, C; Zanella, Riccardo; Zanni, Luca. - In: APPLIED AND COMPUTATIONAL HARMONIC ANALYSIS. - ISSN 1063-5203. - STAMPA. - 27:2(2009), pp. 247-254. [10.1016/j.acha.2009.02.003]
Accelerating gradient projection methods for $ell_1$-constrained signal recovery by steplength selection rules
ZANELLA, RICCARDO;ZANNI, Luca
2009
Abstract
We propose a new gradient projection algorithm that compares favorably with the fastest algorithms available to date for ℓ1-constrained sparse recovery from noisy data, both in the compressed sensing and inverse problem frameworks. The method exploits a line-search along the feasible direction and an adaptive steplength selection based on recent strategies for the alternation of the well-known Barzilai–Borwein rules. The convergence of the proposed approach is discussed and a computational study on both well conditioned and ill-conditioned problems is carried out for performance evaluations in comparison with five other algorithms proposed in the literature.File | Dimensione | Formato | |
---|---|---|---|
0902.4424.pdf
Open access
Tipologia:
Versione dell'autore revisionata e accettata per la pubblicazione
Dimensione
210.48 kB
Formato
Adobe PDF
|
210.48 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris