Research Output per year
My research is related to the study of discrete ill posed linear inverse problems. More specifically, it involves the development of tools for regularisation in the large-scale context, where special attention is given to the use of preconditioning as a regularisation tool. This is based on the construction of relevant models for the problem, linking right and left preconditioning to prior information about the nature of the solution and the noise from the probabilistic model, respectively. Note that, in the case of solution dependent preconditioning, the chosen approach is to exploit information about the solution at each iteration. Therefore, in order to develop solvers for these problems, modified Krylov Subspace methods with regularisation and flexible algorithms with use of preconditioning are investigated. The aim of this project is to exploit the properties of generalised Krylov Subspaces to obtain better reconstructed solutions to discrete high dimensional ill posed linear inverse problems.
Krylov Methods for Inverse Problems: surveying classical, and introducing new, algorithmic approachesGazzola, S. & Sabate Landman, M., 6 Jul 2020, In : Gesellschaft für Angewandte Mathematik und Mechanik (GAMM).
Research output: Contribution to journal › Article