Iterative Regularization and Generalized Discrepancy Principle for Monotone Operator Equations

In this paper, we propose two novel algorithms for solving a nonlinear ill-posed operator equation F(x) = f δ , ||f − f δ || ≤ δ, in a real Hilbert space H under the following basic assumption: The first algorithm is a combination of iteratively regularized Newton's scheme (IRNS) with a posteri...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Numerical functional analysis and optimization 2007-01, Vol.28 (1-2), p.13-25
Hauptverfasser: Bakushinsky, Anatoly, Smirnova, Alexandra
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper, we propose two novel algorithms for solving a nonlinear ill-posed operator equation F(x) = f δ , ||f − f δ || ≤ δ, in a real Hilbert space H under the following basic assumption: The first algorithm is a combination of iteratively regularized Newton's scheme (IRNS) with a posteriori stopping rule The second algorithm is a modified version of IRNS with simultaneous updates of the operator [F′(x n ) + ϵ n I] −1 : combined with ( 1 ). A rigorous theoretical analysis of both methods is conducted.
ISSN:0163-0563
1532-2467
DOI:10.1080/01630560701190315