# Iteration are the stationary values of the same

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: es The eigenvalues of A are the stationary values for x ∈ IRm of the Rayleigh quotient r (x) = xx Ax x The eigenvalue estimates (Ritz values) associated with step n of the Lanczos iteration are the stationary values of the same function r (x) if x is restricted to the Krylov subspace Kn Perfect parallel of what we have shown that the solution x∗ of Ax = b is the minimal point in IRm of the scalar function φ(x), and the CG iterate xn is the minimal point of the same function φ(x) if x is restricted to Kn 10 / 25 Conjugate gradients and polynomial approximation Connection between Krylov subspace iteration and polynomials of matrices The Arnoldi and Lanczos iterations solve the Arnoldi/Lanczos approximation problem Find p n ∈ P n such that p n (A)b = minimum The GMRES iteration solves the GMRES approximation problem Find pn ∈ Pn such that pn (A)b = minimum For CG, the appropriate approximation problem involves the A-norm of the error Find pn ∈ Pn such that pn (A)e0 A = minimum where e0 denotes the initial error e0 = x∗ − x0 = x∗ ,...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online