The paper deals with analytic and numerical studies of linear regularization methods. Thereby Tikhonov regularization and Landweber iteration are compared. In particular the regularization with semi-norms for both methods are considered. A heuristc approach for improving the results for Landweber iteration are introduced. The analytic considerations are illustrated by a detailed numerical study.
Keywords:
ill-posed linear problem, Moore-Penrose inverse, Tikhonov regularization, Landweber iteration, semi-norm, convergence rates, integral equation