A geometric theory for preconditioned inverse iteration. III: A short and sharp convergence estimate for generalized eigenvalue problems (Q1855433)

This paper is a follow-up to two papers by the second author [ibid. 322, 61--85, 87--104 (2001; Zbl 0976.65034, Zbl 0976.65035)]. There a simple preconditioned eigensolver for the computation of the smallest eigenvalue and associated eigenvector of a positive definite matrix was studied and an estimate for its convergence rate was given. In this paper a much shorter and more elegant convergence rate estimate is given. Moreover it is shwon that this estimate holds also for the generalized symmetric eigenvalue problem.