Recursive self preconditioning method based on Schur complement for Toeplitz matrices (Q1944751)

To solve a linear system \(Tx =b\) with a Toeplitz matrix \(T\), this paper describes a recursive self preconditioning algorithm. It is based on repeated halving of the problem and uses the Schur complement and the Gohberg-Semencul matrix inversion formula for computing the generating vectors of \(T^{-1}\). The preconditioned matrices have eigenvalues clustering around 1 and the iterations converge very quickly. Stability is also studied, as are numerical results.