Unformatted text preview: or matrices. We also show that
a matrix and its transpose have the same two norm.
Fact 2.19 (Two Norm). Let A ∈ Cm×n . Then
A∗ A∗ A = A 2, 2 2 = A 2.
2 Proof. The deﬁnition of the two norm implies that for some x ∈ Cn with x
we have A 2 = Ax 2 . The deﬁnition of the vector two norm implies
A 2
2 = Ax 2
2 = x ∗ A∗ Ax ≤ x 2 A∗ Ax 2 2 =1 ≤ A∗ A 2 , where the ﬁrst inequality follows from the Cauchy–Schwarz inequality in Fact 2.8
and the second inequality from the two norm of A∗ A. Hence A 2 ≤ A∗ A 2 .
2
Fact 2.18 implies A∗ A 2 ≤ A∗ 2 A 2 . As a consequence,
A 2
2 ≤ A∗ A 2 ≤ A∗ 2 A 2, A 2 ≤ A∗ 2 . Downloaded 01/17/14 to 143.215.200.123. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php 36 2. Sensitivity, Errors, and Norms The same reasoning applied to AA∗ gives
A∗
Therefore A∗ 2 2
2 ≤ AA∗ =A 2 2 ≤A and A∗ A 2
2 A∗ 2 , A∗ 2 ≤ A 2. = A 2.
2 If we omit a piece of a matrix, the norm does not increase but it can decrease.
Fact 2.20 (Norm of a Submatrix). Let A ∈ Cm×n . If B is a submatrix of A, then
B p ≤ A p. Exercises
(i) Let D ∈ Cn×n be a diagonal matrix with diagonal elements djj . Show that
D p = max1≤j ≤n djj .
(ii) Let A ...
View
Full
Document
This note was uploaded on 01/30/2014 for the course MATH 1502 taught by Professor Mcclain during the Spring '07 term at Georgia Tech.
 Spring '07
 McClain
 Multiplication, Scalar

Click to edit the document details