Convergence of LR algorithm for a one-point spectrum tridiagonal matrix (Q841128)

The paper focuses on the LR algorithm, proving rigorously the convergence of the method applied to an real unreduced tridiagonal matrix with a one-point spectrum - the Jordan form is one big Jordan block. The first section represents an introduction concerning the basic LR algorithm. The second section concerns the eigenvector properties of a one-point spectrum tridiagonal matrix. The authors present some properties of the eigenvalues that are sufficient to guarantee convergence of the basic LR algorithm. Most of the results may be extended directly to complex unreduced tridiagonal matrices. The authors also give an example of a one-point spectrum tridiagonal. The third section focuses on the convergence of basic LR algorithm on a one-point spectrum tridiagonal matrix, taking into consideration the case of an eigenvalue equal to zero and the case of an eigenvalue different to zero.