Preconditioned Inverse Source Solutions

For large-scale inverse source problems with millions of unknowns after discretization, iterative linear systems of equations solvers are the method of choice for the solution of the resulting normal linear systems of equations. In order to reduce the related computational workload, a localized and a global preconditioning technique are presented, which speed up the iterative solution process especially for very large and complex inverse source problems. The performance of the pre-conditioning methods is demonstrated based on synthetic and realistic measurement data.