Scribd
Upload
39 views

S Tutorial 04

This document contains 8 sections about matrix norms and iterative methods: 1) Shows a norm defined on n×n matrices is not subordinate to any vector norm on Rn. 2) Shows the infinity norm of a Vandermonde matrix goes to infinity as n goes to infinity. 3) Shows the condition number of the product of two invertible matrices AB is less than or equal to the product of their individual condition numbers. 4) Estimates the relative error in the solution to a system of equations Ax=b in terms of the relative error in b. 5) Finds the condition number κ2(Tn) of a tri-diagonal matrix Tn. 6) Bounds the infinity

Uploaded by

Raghav Jain
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
39 views

S Tutorial 04

This document contains 8 sections about matrix norms and iterative methods: 1) Shows a norm defined on n×n matrices is not subordinate to any vector norm on Rn. 2) Shows the infinity norm of a Vandermonde matrix goes to infinity as n goes to infinity. 3) Shows the condition number of the product of two invertible matrices AB is less than or equal to the product of their individual condition numbers. 4) Estimates the relative error in the solution to a system of equations Ax=b in terms of the relative error in b. 5) Finds the condition number κ2(Tn) of a tri-diagonal matrix Tn. 6) Bounds the infinity

Uploaded by

Raghav Jain
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Section

Tutorial Sheet 4
Matrix Norms and Iterative Methods
1. Show that the norm defined on the set of all n × n matrices by
n !
! n
"A" = |aij |.
i=1 j=1

is not subordinate to any vector norm on Rn .

2. For a given vector (x1 , . . . , xn )T , the corresponding Vandermonde matrix is de-


fined as
Vn = (vij ) 1≤i≤n
1≤j≤n

where vij = xj−1


i . For a fixed integer n > 1, let xk = 1 + k/n, for k = 1, 2, . . . , n,
then show that "Vn "∞ → ∞ as n → ∞.

3. Let A and B be invertible matrices with condition numbers κ(A) and κ(B) re-
spectively. Show that κ(AB) ≤ κ(A)κ(B).

4. In solving the system of equations Ax = b with matrix


" #
1 4
A= ,
1 4.01

estimate the relative error in the solution vector x in terms of the relative error
in b. Test your estimate in the case when b = (4, 4)T and b̃ = (3.95, 4.01)T . Use
the maximum norm for vectors in R2 .
5. For n ≥ 2, find κ2 (Tn ) where Tn is the tri-diagonal matrix given by
 
2 −1 0

 −1
 . .
2 −1
. . . .


 0
Tn =  . . . 


 0 −1 2
0 −1
−1
2

1
S. Sivaji Ganesh and S. Baskar Spring 2022-23
Section

6. Let x(12) be the 12th term of the Gauss-Seidel iterative sequence for the system

3x1 + 2x2 = 1
4x1 + 12x2 + 3x3 = −2
x1 + 3x2 − 5x3 = 3

with x(0) = (0, 0, 0)T . If x denotes the exact solution of the given system, then
show that
"e(12) "∞ ≤ 0.0077073467"x"∞ .

7. Consider the linear system Ax = b, where A = (aij )1≤i,j≤n ∈ Mn (R) be such that
aii )= 0 for i = 1, 2, . . . , n, and b = (b1 , b2 , . . . , bn )T ∈ Rn . For a given real number
ω )= 0, define an iterative sequence {x(k) } as
* i−1 n
+
1 ! !
Given x(0) ; zi
(k+1) (k+1) (k)
= bi − aij xj − aij xj ,
aii j=1 j=i+1

i = 1, 2, · · · , n
, for
(k+1) (k+1)
xi = (1 − ω)xki + ωzi
k = 0, 1, 2, . . . .

This iterative method is called the successive over relaxation method (SOR method).
Find the matrix Sω and the vector c such that the SOR iterative method is written
in the matrix form

x(k+1) = Sω x(k) + c, k = 0, 1, 2, . . . .

8. Let A be an n × n matrix with real entries. Let κ2 (A) and κ∞ (A) denote the
condition numbers of a matrix A that are computed using the matrix norms "A"2
and "A"∞ , respectively. Answer the following questions.
i) Determine all the diagonal matrices such that κ∞ (A) = 1.
ii) Let Q be a matrix such that QT Q = I (such matrices are called orthogonal
matrices). Show that κ2 (Q) = 1.
iii) If κ2 (A) = 1, show that all the eigenvalues of AT A are equal. Further, de-
duce that A is a scalar multiple of an orthogonal matrix.

2
S. Sivaji Ganesh and S. Baskar Spring 2022-23

You might also like