0% found this document useful (0 votes)
18 views10 pages

5.0 Iterative - Methods

Uploaded by

Kiryei Takashi
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
0% found this document useful (0 votes)
18 views10 pages

5.0 Iterative - Methods

Uploaded by

Kiryei Takashi
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/ 10

College of Engineering

Civil Engineering Department

ITERATIVE METHODS
• Introduction

Contents
• Gauss – Seidel Method
• Jacobi Method
• In solving sets of linear algebraic equations by
elimination, iterative methods are preferred over
direct methods especially when the coefficient
INTRODUCTION

matrix is sparse (has many zeros).


• Iterative Methods may converge rapidly and may
be more economical in memory requirement of
computers.
• For hand computations, the have the distinct
advantage of being self – correcting if errors are
encountered in the solution.
• They are sometimes used to reduce round – off
errors in the solution computed by Direct Methods.
• Iterative Methods are also applied to find
solutions for nonlinear equations.
➢ One of the most useful technique in
solving large numbers of simultaneous
equations.
GAUSS - SEIDEL

➢ Gauss – Seidel Method may not always


or may sometimes slowly converge to a
solution.
➢ This method will always converge to a
solution when the set of equations are
arranged in a “diagonal system”.
➢ A “diagonal system is sufficient to
ensure convergence but is not
necessary.
Algorithm for Gauss – Seidel

Given is a set of simultaneous linear algebraic equations:


GAUSS - SEIDEL

a11X1 + a12X2 + a13X3 = C1


a21X1 + a22X2 + a23X3 = C2
a31X1 + a32X2 + a33X3 = C3

1. Assign initial value for each unknown starting with the


first equation, solve for the largest unknown coefficient
say:
X2 = 0; X3 = 0; X1 = ?
a11X1 + a12(0) + a13(0) = C1

X1 = C1/a11
2. Go to the 2nd equation, solve for the largest coefficient.
X1 = C1/a11; X3 = 0; X2 = ?

a21X1 + a22(X2) + a23(0) = C2


GAUSS - SEIDEL
a21(C1/a11 )+ a22(X2) + a23(0) = C2

C2−a21(C1/a11 )
X2 =
a22

3. Go to the 3rd equation and solve for the largest coefficient.


C −a (C /a )
X1 = C1/a11; X2 = 2 21 1 11 ; X3 = ?
a22

a31X1 + a32(X2) + a33(X3) = C3


C −a (C /a )
a31 (C1/a11 ) + a32{ 2 21 1 11 } + a33(X3) = C3
a22

C C −a (C /a )
X3 = {C3 – [a31 ( 1 ) + a32 [ 2 21 1 11 ]}/a3𝟑
a11 a22
4. Solve for the new Xi
5. Continue iteration until the value of each unknown
determined in a particular iteration differs from its respective
value obtained in the preceding iteration by an amount less
than some arbitrary selected approximate error / σ𝑎/.
GAUSS - SEIDEL

𝑃𝑟𝑒𝑠𝑒𝑛𝑡 −𝑃𝑟𝑒𝑣𝑖𝑜𝑢𝑠
Where: / σ𝑎/ = / / x 100
𝑃𝑟𝑒𝑠𝑒𝑛𝑡

/ σ𝑎/≤ / σ𝑠/
➢ Jacobi Method is also known as the
Method of Simultaneous Displacement.
JACOBI METHOD

➢ Each of the equations is simultaneously


changed by using the most recent set
of variables.
➢ Just like Gauss – Seidel Method, Jacobi
iteration follows the “diagonal system”
to ensure convergence to a solution.
Systems of Linear Equations Algorithm for Jacobi Method

Given is a set of simultaneous linear algebraic equations:

a11X1 + a12X2 + a13X3 = C1


a21X1 + a22X2 + a23X3 = C2
a31X1 + a32X2 + a33X3 = C3

Rewrite the equations in the following form:

(𝒏+𝟏) 𝑪𝟏−𝒂𝟏𝟐𝒙𝟐𝒏−𝒂𝟏𝟑𝒙𝟑𝒏
𝑿𝟏 =
𝒂𝟏𝟏

(𝒏+𝟏)
𝑪𝟐 − 𝒂𝟐𝟏𝒙𝟏𝒏 − 𝒂𝟐𝟑𝒙𝟑𝒏
𝑿𝟐 =
𝒂𝟐𝟐

(𝒏+𝟏)
𝑪𝟑 − 𝒂𝟑𝟏𝒙𝟏𝒏 − 𝒂𝟑𝟐𝒙𝟐𝒏
𝑿𝟑 =
𝒂𝟑𝟑
Systems of Linear Equations
2. 1.

1. Assume initial approximations to the value of


variables X1n, X2n and X3n.
2. Substitute these initial approximations on the right -
hand side of the set of equations (re-written
equations) to get new approximations - X1n+1, X2n+1
and X3n+1.
3. The process is repeated until successive values of
each of the variables are sufficiently alike or if the
value of each unknown determined in a particular
iteration differs from its respective value obtained in
the preceding iteration by an amount less than the
arbitrary selected approximate error / σ𝑎/.

/ σ𝑎/≤ / σ𝑠/

You might also like