Scribd
Upload
11 views

MAFE208IU L5 - Linear Systems of Equations - Part 2

The document discusses the Gauss-Seidel method for solving systems of linear equations. It explains how to transform the equations and iteratively solve for the unknowns until convergence within a specified tolerance is reached. An example applying the method to a 3x3 system is provided and the advantages of Gauss-Seidel for sparse matrices are noted.

Uploaded by

uph.huyen
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
11 views

MAFE208IU L5 - Linear Systems of Equations - Part 2

The document discusses the Gauss-Seidel method for solving systems of linear equations. It explains how to transform the equations and iteratively solve for the unknowns until convergence within a specified tolerance is reached. An example applying the method to a 3x3 system is provided and the advantages of Gauss-Seidel for sparse matrices are noted.

Uploaded by

uph.huyen
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/ 13

Chapter 2: Solutions of Linear

Systems of Equations
Lecture 2
Gauss-Seidel iteration Method
Gauss-Seidel Method
A set of n equations and n unknowns:
a11 x1  a12 x2  a13 x3  ...  a1n xn  b1
a21 x1  a22 x2  a23 x3  ...  a2n xn  b2
……
an1x1  an 2 x2  an3 x3  ...  ann xn  bn

The diagonal elements are non-zero

Rewrite each equation solving for the corresponding


unknown. Ex:
• First equation, solve for x1
• Second equation, solve for x2
MAFE208IU-L5 • …. 2
Gauss-Seidel Method
Transformation
Rewriting each equation
b1  a12 x2  a13 x3   a1n xn
x1 
a11 From Equation 1

b2  a21 x1  a23 x3   a2 n xn
x2  From equation 2
a22
…..
  
bn 1  an 1,1 x1  an 1,2 x2   an 1,n  2 xn  2  an 1,n xn
xn 1  From equation n-1
an 1,n 1
bn  an1 x1  an 2 x2    an,n 1 xn 1
xn  From equation n
ann
MAFE208IU-L5 3
Algorithm for Gauss-Seidel Method
Let x ( k )  ( x1( k ) , x2( k ) ,...., xn( k ) ) be a given approximate value,
k  0,1, 2,..., we compute the new approximate value x ( k 1) :

b1  a12 x2( k )  a13 x3( k )  a1n xn( k )


x1( k 1) 
a11

bi  ai1 x1( k 1)  ...  ai ,i 1 xi(k11)  ai ,i 1 xi(k1)  a1n xn( k )


xi( k 1) 
aii

bn  an1 x1( k 1)  ...  a1,n 1 xn( k11)


xn( k 1) 
ann
Repeat this process until a stopping criterion is met
MAFE208IU-L5 4
Stopping Criteria
We can use either Approximate Error or Relative
Approximate Error to stop the computation process

Ea  max{| x new
i x old
i |, i  1, 2,..., n}
new
x x old
 a  max{ i
new
i
, i  1, 2,..., n}
xi

The iterations are stopped when the absolute


relative approximate error is less than a
prespecified tolerance for all unknowns x1 , x2 ,..., xn
.
MAFE208IU-L5 5
Example
Use Gauss-Seidel method, find the root of the system with stopping
criterion |Ea | ≤ 0.01 with starting point x(0) =(0,0,0):
6 x1  x2  2 x3  1
x1  8 x2  3x3  5
2 x1  2 x2  9 x3  7
Solution
Resolve for each variable x1  (1  x2  2 x3 ) / 6
x2  (5  x1  3x3 ) / (8)
x3  (7  2 x1  2 x2 ) / (9)
Iterative Formula:
x1( k 1)  (1  x2( k )  2 x3( k ) ) / 6
x2( k 1)  ( 5  x1( k 1)  3x3( k ) ) / ( 8)
MAFE208IU-L5 x3( k 1)  ( 7  2 x1( k 1)  2 x2( k 1) ) / ( 9) 6
Example …
Starting with x(0) =(0,0,0), and using:

x1( k 1)  (1  x2( k )  2 x3( k ) ) / 6


( k 1) ( k 1)
x 2  ( 5  x 1  3x ) / ( 8)
(k )
3

x3( k 1)  ( 7  2 x1( k 1)  2 x2( k 1) ) / ( 9)


we get
(k )
x x1 | x1new  x1old | x2 | x2new  x2old | x3 | x3new  x3old | Ea
0 O 0 O
1 0.1667 0.6458 0.8843 --
2 -0.0204 0.1871 0.9540 0.3082 0.9943 0.1100 0.3082
3 -0.0058 0.0146 0.9972 0.0432 1.0006 0.0063 0.0432
4 -0.0007 0.0051 1.0002 0.0030 1.0002 0.0004 0.0051

So, approximate solution is x=(-0.0007, 1.0002, 1.0002)


MAFE208IU-L5 7
Remark
Gauss-Seidel method: not all systems of equations
will converge.
One class of system of equations always converges:
One with a diagonally dominant coefficient matrix.
Diagonally dominant: [A] in [A] [X] = [C] is diagonally
dominant if:
n n
aii   aij
for all ‘i’ aii   aij
j 1 j 1
for at least one ‘i’
j i and j i

MAFE208IU-L5 8
Comments on Gauss-Seidel
• Disadvantages:
– it may not converge
– When it converged, it did so very slowly
• Advantages:
– When matrix is very sparse (most elements are
zero), elimination methods waste large amount of
computer memory by storing zeros
– Gauss-Seidel method saves memory because only
nonzero coefficients are involved in the structure
of equations
MAFE208IU-L5 9
Exercise
Using Gauss-Seidel iteration method, solve the following system

3 x1  x2  5
2 x2  x3  1
x2  5 x3  2 x4  25
x3  6 x4  x5  16
 x4  3x5  11
Use x (0) =(0,0,0,0,1), stopping condition: | Ea | 0.02

MAFE208IU-L5 10
Quiz 2
2) Using Gauss-Seidel
mn: last two digits of your iteration method, solve the
Student ID number following system

1) Solve the following (2  3m) x1  x3  3m  5


system using Crout’s
method x1  (1  2n) x2  4n  1
x2  (4  n) x3  2 x4  3n  22
2 x1  nx2  3x3  2n  11
x2  5 x4  x5  27
nx1  (3  m) x2  3 x3  2m  n  15
mx1  x2  2 x3  m  4 2 x3  (3  m) x5  5m  9

Use x (0) =(0,0,0,0,0),


stopping condition: | Ea | 0.02
MAFE208IU-L5 11
Ex 2: Find temperature in a square sheet of metal. The temperature
at the edges of the sheet are kept at : 100, 90, 50, and 0 degrees.
T1,3  90 T  90 T3,3  90
2,3

T1, 2 T2, 2 T3,2


T0,2  100 T4,2  50

T1,1 T2,1 T3,1


T0,1  100 T4,1  50

T1,0  0 T2,0  0 T3,0  0


Temperature at grid points: Ti 1, j  Ti 1, j  Ti , j 1  Ti , j 1  4Ti , j  0
MAFE208IU-L14 Solve by Gauss-Seidel method, use |𝐸 | ≤ 0.2 12
T1,3  200 T2,3  200
Quiz
Find temperature in a
square sheet of metal.
T0,2  10m T1, 2 T2, 2 T3,2  10n
The temperature at
edges of the sheet are
kept as in Figure T0,1  5m T3,1  5n
T1,1 T2,1
Approximate T
at grid points
Temperature at grid points:
T1,0  0 T2,0  0
Ti 1, j  Ti 1, j  Ti , j 1  Ti , j 1  4Ti , j  0
mn  The last two digits
Solve by a) Gauss elimination of your student ID number
b) LU decomposition method
c) Gauss-Seidel method, use |𝐸 | ≤ 0.5
MAFE208IU-L14 13

You might also like