Volume 16, pp. 107-128, 2003.

A parameter choice method for Tikhonov regularization

Limin Wu

Abstract

A new parameter choice method for Tikhonov regularization of discrete ill-posed problems is presented. Some of the regularized solutions of a discrete ill-posed problem are less sensitive than others to the perturbations in the right-hand side vector. This method chooses one of the insensitive regularized solutions using a certain criterion. Numerical experiments show that the new method is competitive with the popular L-curve method. An analysis of the new method is given for a model problem, which explains how this method works.

Full Text (PDF) [226 KB]

Key words

discrete ill-posed problems, discrete Picard condition, Tikhonov regularization.

AMS subject classifications

65F22.

< Back