508 insures ann utu for a unitary u and an upper

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: . , λk } , then the spectral projector onto N (A − λi I) along R (A − λi I) is given by k k (A − λj I) Gi = O C j =1 j =i (λi − λj ) for i = 1, 2, . . . , k. (7.3.11) j =1 j =i Consequently, if f (z ) is deﬁned on σ (A) , then f (A) = is a polynomial in A of degree at most k − 1. k i=1 f (λi )Gi Problem: For a scalar t, determine the matrix exponential eAt , where A= −α α β −β with α + β = 0. Solution 1: The characteristic equation for A is λ2 + (α + β )λ = 0, so the eigenvalues of A are λ1 = 0 and λ2 = −(α + β ). Note that A is diagonalizable Copyright c 2000 SIAM Buy online from SIAM http://www.ec-securehost.com/SIAM/ot71.html Buy from AMAZON.com 530 Chapter 7 Eigenvalues and Eigenvectors http://www.amazon.com/exec/obidos/ASIN/0898714540 because no eigenvalue is repeated—recall (7.2.6). Using the function f (z ) = ezt , the spectral representation (7.3.6) says that eAt = f (A) = f (λ1 )G1 + f (λ2 )G2 = eλ1 t G1 + eλ2 t G2 . It is illegal to print, duplicate, or distribute this material Please report violations to [email protected] The spectral projectors G1 and G2 are determined from (7.3.11) to be G1 = A − λ2 I 1 = −λ2 α+β β α β α and G2 = A 1 = λ2 α+β α −α 1 α+β β α β α + e−(α+β )t α −α , D E so eAt = G1 + e−(α+β )t G2 = −β β T H −β β . Solution 2: Compute eigenpairs (λ1 , x1 ) and (λ2 , x2 ), construct P = x1 | x2 , and compute IG R 0 f (λ1 ) 0 f (λ2 ) eAt = P eλ1 t 0 P−1 = P 0 eλ2 t P−1 . The computational details are called for in Exercise 7.3.2. Example 7.3.4 Y P Problem: For T = 1 /2 1 /4 1 /2 3 /4 , evaluate limk→∞ Tk . Solution 1: Compute two eigenpairs, λ1 = 1, x1 = (1, 1)T , and λ2 = 1/4, x2 = (−2, 1)T . If P = [x1 | x2 ], then T = P 1 104 P−1 , so 0 / O C Tk = P 1k 0 0 1/4k P−1 → P 1 0 0 0 P−1 = 1 3 1 1 2 2 . (7.3.12) Solution 2: We know from (7.3.6) that Tk = 1k G1 + (1/4)k G2 → G1 . Since λ1 = 1 is a simple eigenvalue, formula (7.2.12) on p. 518 can be used to compute T T T G1 = x1 y1 /y1 x1 , where x1 and y1 are any right- and left-hand eigenvectors associated with λ1 = 1. A right-hand e...
View Full Document

This document was uploaded on 03/06/2014 for the course MA 5623 at City University of Hong Kong.

Ask a homework question - tutors are online