A wavelet multilevel method for ill-posed problems stabilized by Tikhonov regularization (Q1358140)

The paper is concerned with the construction of a multilevel solver for a compact operator equation of the first kind. The equation is discretized by applying the method of least squares and then it is stabilized by the Tikhonov regularization algorithm. To achieve an efficient iterative solver in terms of convergence speed, parallel coding and performance, the author considers an additive Schwarz relaxation method. For the construction of the multilevel solver, the test functions space is splitted into orthogonal subspaces of increasing dimensions. Wavelets are already used for the treatment of an inverse problem. The performace of the proposed algorithm is checked for integral equations of the first kind.