Lecture #9 Notes

# 510 theorem let a mn with eigenvalues 1 2 n

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: 1 02 01 10 A1 = = = = Q 1 R1 01 10 10 10 02 ￿ ￿￿ ￿￿ ￿ 10 01 01 A2 = = = A0 02 10 20 So from here on, the cycle repeats. This leads to repeated A￿j s which fails to converge. Q: What conditions ensure convergence of the QR algorithm? 2.5.10 Theorem. Let A ∈ Mn with eigenvalues λ1 , λ2 , . . . , λn (multiplicity counted) and |λj | = ￿ |λk | for j ￿= k . Then the QR algorithm converges • The rate at which it converges deals with the relation between |λj | and |λk | Reality Check: What about matrices with real entries? What works and what does not? 2.5...
View Full Document

## This note was uploaded on 01/16/2014 for the course MATH 6304 taught by Professor Bernhardbodmann during the Fall '12 term at University of Houston.

Ask a homework question - tutors are online