Recent Developments in Inverse Problems - Abstract

Clason, Christian

Convex regularization of hybrid discrete-continuous inverse problems

We consider linear and nonlinear parameter identification problems, where the unknown distributed parameter is known to take on values only from a discrete given set. This property can be promoted with the aid of a convex penalty. A regularized semismooth Newton method allows the numerical solution of the corresponding Tikhonov functional. Numerical examples illustrate the effectiveness of the proposed approach.