Scribd
Upload
89 views2 pages

MATH3230A Numerical Analysis: Tutorial 9

This document discusses numerical analysis concepts including: 1. The LU factorization of a tridiagonal matrix and solving a linear system using this factorization. 2. Single and double precision floating point number representations in binary. Rounding errors in floating point arithmetic are estimated. 3. Deriving an algorithm to solve linear systems with tridiagonal matrices by eliminating terms through successive rows.

Uploaded by

RattanLal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
89 views2 pages

MATH3230A Numerical Analysis: Tutorial 9

This document discusses numerical analysis concepts including: 1. The LU factorization of a tridiagonal matrix and solving a linear system using this factorization. 2. Single and double precision floating point number representations in binary. Rounding errors in floating point arithmetic are estimated. 3. Deriving an algorithm to solve linear systems with tridiagonal matrices by eliminating terms through successive rows.

Uploaded by

RattanLal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

MATH3230A Numerical Analysis

Tutorial 9
1. Consider the following tridiagonal matrix

b1
a2

A=

c1
b2

..

c2

..

..

an1

bn1
an

cn1
bn

(a) What is the LU factorization of A?


(b) Using the result of (a), solve the following linear system

4
1

1
4
1

1
4
1

1
4
1

1
4

x1
x2
x3
x4
x5

2
6
0
6
2

2. Recall that most computers adopt the binary system. A machine number is a string consisting of bits, whose
value are decoded as the following normalized oating point:
p

a = (1)s q 2(1)

For a 32-bits oating-point binary storage formats (also called single precision), we have
1 bit for the sign (s), 8 bit for the exponent (1 bit for p,7 bits for m), 23 bits for the mantissa.
For a 64-bits oating-point binary storage formats (also called double precision), we have
1 bit for the sign (s), 11 bits for the exponent (1 bit for p,10 bits for m), 52 bits for the mantissa.
During the decimal-binary conversion, small roundo errors occurs.
Also rounding is usually adapted in scientic computing. If x is rounded to xe with n digits after the decimal
points, we have the error estimate
|x x
e|

1
10n
2

In addition to rounding input, rounding is also needed after most arithmetic operations. The roundo error
is less than 224 (32-bits) or 253 (64-bits) and it is called the unit rounded error. All these put together to
form oating point arithmetics.
Given a real number x, let f l(x) be the oating point representation of x, which means


f l(x) x
2 := m



x

where m is the machine precision/ machine unit roundo error. Then we can write
f l(x) = x(1 + )

with  m .
Estimate the approximation errors for the following oating point operations. You can use  to represent
the machine precision.
1

(a)
(b)
(c)
(d)
(e)

an , where n is a positive integer


(a + b)(a b)
xT x, where x Rn is a n dimensional column vector.
(a2 + b2 c)d
a2 b2 c

3. (Challenging Question) Let A = (aij ) Rnn be strictly diagonally dominant matrix, i.e.,
n
X

|akk | >

|akj | ,

k = 1, 2, ...., n

j=1,j6=k

(a) Prove that a strictly diagonally dominant matrix is nonsingular.


(b) Prove that if the strictly diagonally dominant matrix is symmetric and with positive diagonal entries,
then it is positive denite.
4. (Challenging Question) In this question, we are going to derive another algorithm to solve 3(b).
(a) Note that the following linear system

b1
a2

c1
b2

..

c2

..

an1

..

bn1
an

cn1
bn

x1
x2
x3
x4
x5

y1
y2
y3
y4
y5

can be written as
ai xi1 + bi xi + ci xi+1 = yi ,

i = 1, 2, ..., n,

a1 = cn = 0

Conisder the rst two row of equations. Try to eliminiate the term containing x1 .
(b) From (a), we obtained the modied second equation. Try to use the modied second equation and the
third equation to eliminate the term containing x2 and obtain
(b3 (b1 b2 c1 a2 ) c2 b1 a3 ) x3 + c3 (b1 b2 c1 a2 )x4

= y3 (b1 b2 c1 a2 ) (y2 b1 y1 a2 ) a3

(c) Using (a) and (b), try to formulate an algorithm for solving tridiagnoal matrix.
(d) Using the above algorithm, try to solve 5(b) again.
(e) What is the computational cost of the algorthim derived above?

You might also like