Some theoretical comparisons of refined Ritz vectors and Ritz vectors

Refined projection methods proposed by the author have received attention internationally. We are concerned with a conventional projection method and its refined counterpart for computing approximations to a simple eigenpair (A, x) of a large matrix A. Given a subspace ω that contains an approximation to x, these two methods compute approximations (μ,x) and (μ,x) to (λ, x), respectively. We establish three results. First, the refined eigenvector approximation or simply the refined Ritz vector x is unique as the

Some theoretical comparisons of refined Ritz vectors and Ritz vectors

Refined projection methods proposed by the author have received attention internationally. We are concerned with a conventional projection method and its refined counterpart for computing approximations to a simple eigenpair (λ,x)of a large matrix A. Given a subspace w that contains anapproximation to x, these two methods compute approximations(μ(x~)) and (μ(x^)) to (λ,x),respectively. We establish three results. First, the refinedeigenvector approximation or simply the refined Ritz vector (x^) is unique as the deviation of x from w approaches zero if λ is simple. Second, interms of residual norm of the refined approximate eigenpair (μ,(x^)), we derive lower and upper bounds for the sine of the angle betweenthe Ritz vector (x~) and the refined eigenvector approximation (x^), and we prove that (x~)≠(x^) unless (x^)=x. Third, we establish relationships between theresidual norm ‖A(x~)-μ(x^)‖ of the conventionalmethods and the residual norm ‖A(x^)-μ(x^)‖ of therefined methods, and we show that the latter is always smallerthan the former if (μ,(x^)) is not an exact eigenpair ofA, indicating that the refined projection method is superiorto the corresponding conventional counterpart.