A geometric convergence theory for the preconditioned steepest descent iteration
Provides a theoretical convergence guarantee for a widely used eigensolver, enabling practical use of preconditioners without ideal scaling.
The paper derives a sharp non-asymptotic convergence estimate for the preconditioned steepest descent iteration for eigenvalue problems, showing it always improves upon the fixed-step iteration and allows arbitrarily scaled preconditioners.
Preconditioned gradient iterations for very large eigenvalue problems are efficient solvers with growing popularity. However, only for the simplest preconditioned eigensolver, namely the preconditioned gradient iteration (or preconditioned inverse iteration) with fixed step size, sharp non-asymptotic convergence estimates are known and these estimates require an ideally scaled preconditioner. In this paper a new sharp convergence estimate is derived for the preconditioned steepest descent iteration which combines the preconditioned gradient iteration with the Rayleigh-Ritz procedure for optimal line search convergence acceleration. The new estimate always improves that of the fixed step size iteration. The practical importance of this new estimate is that arbitrarily scaled preconditioners can be used. The Rayleigh-Ritz procedure implicitly computes the optimal scaling.