Volume 47, pp. 73-99, 2017.

On generalized iterated Tikhonov regularization with operator-dependent seminorms

Davide Bianchi and Marco Donatelli


We investigate the recently introduced Tikhonov regularization filters with penalty terms having seminorms that depend on the operator itself. Exploiting the singular value decomposition of the operator, we provide optimal order conditions, smoothing properties, and a general condition (with a minor condition of the seminorm) for the saturation level. Moreover, we introduce and analyze both stationary and nonstationary iterative counterparts of the generalized Tikhonov method with operator-dependent seminorms. We establish their convergence rate under conditions affecting only the iteration parameters, proving that they overcome the saturation result. Finally, some selected numerical results confirm the effectiveness of the proposed regularization filters.

Full Text (PDF) [1.2 MB], BibTeX

Key words

ill-posed problems, fractional Tikhonov regularization, iterated Tikhonov, filter functions

AMS subject classifications

65F22, 47A52, 65R32

Links to the cited ETNA articles

[6]Vol. 29 (2007-2008), pp. 163-177 Marco Donatelli and Stefano Serra-Capizzano: Filter factor analysis of an iterative multilevel regularizing method

ETNA articles which cite this article

Vol. 55 (2022), pp. 169-186 Davide Bianchi, Alessandro Buccini, Marco Donatelli, and Emma Randazzo: Graph Laplacian for image deblurring

< Back