Numerical Methods - Assignment 5 | MATH 417, Assignments of Mathematical Methods for Numerical Analysis and Optimization

Material Type: Assignment; Class: NUMERICAL METHODS; Subject: MATHEMATICS; University: Texas A&M University; Term: Unknown 1989;

Typology: Assignments

Pre 2010

Uploaded on 02/10/2009

koofers-user-4hz
koofers-user-4hz 🇺🇸

10 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
MATH 417: Numerical Analysis
Instructor: Prof. Wolfgang Bangerth
Teaching Assistants: Dukjin Nam
Homework assignment 5 due 3/8/2007
Problem 1 (LU decomposition). Write a program that implements the LU
decomposition algorithm for general n×nmatrices and outputs the Land U
factors. Apply it to the matrix of the linear system
11
2
1
3
1
4
1
2
1
3
1
4
1
5
1
3
1
4
1
5
1
6
1
4
1
5
1
6
1
7
x1
x2
x3
x4
=
1
2
3
4
.
In a second step, implement the backward and forward substitution solves
with the upper and lower triangular factors Land Ufor any given vector. Apply
it to the right hand side above and verify that your solution is correct.
(6 points)
Problem 2 (Norms on Rn). In the analysis of iterative solution methods
for linear systems, we will come across different vector norms. A functional
k·k:RnRis called a norm if it satisfies the following three conditions:
(a) kxk 0 for all vectors xRnand kxk= 0 if and only if x= 0 (positive
definiteness);
(b) kλxk=|λ|kxkfor all λRand all vectors xRn(scalability);
(c) kx+yk kxk+kykfor all vectors x, y Rn(triangle inequality).
Determine which of the following are norms on Rnby proving or disproving
that they satisfy the three conditions above:
a) max1in|xi|
b) max2in|xi|
c) Pn
i=1 |xi|3
d) Pn
i=1 |xi|1/22
e) max{|x1x2|,|x1+x2|,|x3|,|x4|, . . . , |xn|}
f) Pn
i=1 2i|xi|(6 points)
(please turn over)
1
pf2

Partial preview of the text

Download Numerical Methods - Assignment 5 | MATH 417 and more Assignments Mathematical Methods for Numerical Analysis and Optimization in PDF only on Docsity!

MATH 417: Numerical Analysis

Instructor: Prof. Wolfgang Bangerth

[email protected],

Teaching Assistants: Dukjin Nam

[email protected]

Homework assignment 5 – due 3/8/

Problem 1 (LU decomposition). Write a program that implements the LU

decomposition algorithm for general n × n matrices and outputs the L and U

factors. Apply it to the matrix of the linear system

1 2

1 3

1 4 1 2

1 3

1 4

1 5 1 3

1 4

1 5

1 6 1 4

1 5

1 6

1 7

x 1

x 2

x 3

x 4

In a second step, implement the backward and forward substitution solves

with the upper and lower triangular factors L and U for any given vector. Apply

it to the right hand side above and verify that your solution is correct.

(6 points)

Problem 2 (Norms on R

n ). In the analysis of iterative solution methods

for linear systems, we will come across different vector norms. A functional

‖ · ‖ : R

n → R is called a norm if it satisfies the following three conditions:

(a) ‖x‖ ≥ 0 for all vectors x ∈ R n and ‖x‖ = 0 if and only if x = 0 (positive

definiteness);

(b) ‖λx‖ = |λ|‖x‖ for all λ ∈ R and all vectors x ∈ R

n (scalability);

(c) ‖x + y‖ ≤ ‖x‖ + ‖y‖ for all vectors x, y ∈ R

n (triangle inequality).

Determine which of the following are norms on R

n by proving or disproving

that they satisfy the three conditions above:

a) max 1 ≤i≤n |xi|

b) max 2 ≤i≤n |xi|

c)

n i= |xi|

3

d)

n i= |xi|

1 / 2

e) max{|x 1 − x 2 |, |x 1 + x 2 |, |x 3 |, |x 4 |,... , |xn|}

f)

∑n

i=

−i |xi| (6 points)

(please turn over)

Problem 3 (Vector and matrix norms). Compute the norms ‖.‖∞ and

‖.‖ 2 for

x 1 =(3, − 4 , 0 , 1 .5)

T

x 2 =(2, 1 , − 3 , 4)

T

x 3 =(sin k, cos k, 2

k )

T for k ∈ N

x 4 =

4 k+

2 k^2 , k

2 e

−k

)T

for k ∈ N

Compute the row-sum norm ‖.‖∞ of the following matrices:

A =

B =

C =

 D =

(4 points)

Problem 4 (Equivalence of norms on R

n ). In class, we proved the equiv-

alence of the norms ‖.‖∞ and ‖.‖ 2. Here now, prove the same for ‖.‖∞ and ‖.‖ 1 ,

where

‖x‖ 1 =

n ∑

i=

|xi|.

a) Prove that there are indeed constants c, C such that

c‖v‖∞ ≤ ‖v‖ 1 ≤ C‖v‖∞.

where

‖v‖ 1 =

i

|vi|,

‖v‖∞ = max i

|vi|,

and where v is an n-dimensional vector in R

n .

b) For vectors v 1 , v 2 with ‖v 1 ‖ 1 ≤ ‖v 2 ‖ 1 , does the result of part a) imply that

‖v 1 ‖∞ ≤ ‖v 2 ‖∞? If not, give an example of vectors for which this does

not follow. (4 points)