Zur Seitenansicht
 

Titelaufnahme

Titel
Comparison of Landweber Iteration and Iteratively Regularized Gauss-Newton Method / submitted by David Benjamin Blüher, BSc
AutorInnenBlüher, David Benjamin
Beurteiler / BeurteilerinNeubauer, Andreas
ErschienenLinz, 2017
Umfang60 Blätter : Illustrationen
HochschulschriftUniversität Linz, Masterarbeit, 2017
SpracheEnglisch
DokumenttypMasterarbeit
Schlagwörter (DE)Landweber / Gauss-Newton / inverse Probleme
Schlagwörter (EN)Landweber / Gauss-Newton / inverse problems
Schlagwörter (GND)Iteration / Gauß-Newton-Methode / Inverses Problem
URNurn:nbn:at:at-ubl:1-18902 Persistent Identifier (URN)
Zugriffsbeschränkung
 Das Werk ist gemäß den "Hinweisen für BenützerInnen" verfügbar
Dateien
Comparison of Landweber Iteration and Iteratively Regularized Gauss-Newton Method [0.46 mb]
Links
Nachweis
Klassifikation
Zusammenfassung (Englisch)

This thesis deals with two iterative methods that are used to solve nonlinear ill-posed problems, the Landweber iteration, and the iteratively regularized Gauss-Newton method. Both methods are introduced theoretically and the main results on convergence and convergence rates that were established in recent times are reviewed. In order to prove these results, in either case, very similar assumptions are required which makes it easier to compare them. The second part of the work is concerned with the numerical implementation and comparison of results for different integral operators of the Hammerstein type. The focus therein is on showing that the necessary assumptions are fulfilled and that the numerical tests agree with the theoretical results, as well as comparing convergence rates and CPU time between the Landweber iteration and the iteratively regularized Gauss-Newton method. We will, however, also see cases in which the theoretical assumptions that are required for the convergence rates are not fulfilled or cannot be shown, but where an appropriate solution and even rates can be achieved by the same algorithms.

Statistik
Das PDF-Dokument wurde 31 mal heruntergeladen.