Inverse Toeplitz preconditioners for ill-posed problems (Q1124768)

The authors establish a relationship between the preconditioned system and the prolate matrix. A Toeplitz approximate inverse preconditioning scheme for discrete ill-posed problems is considered and it is proved that the preconditioned system approximate the prolate matrix, and that this property implies fast convergence of conjugate gradient type methods. One- and two-dimensional cases are considered. An image restoration application is used to demonstrate the properties of the above preconditioner. The numerical results reported in the paper verified the good theoretical convergence properties of the new preconditioner for a realistic problem.