Numerical Methods

Week 1
 Introduction

Numerical analysis techniques available to solve a

range of mathematical problems encountered in
engineering. Numerical analysis has numerous
applications in all fields of science and some fields
of engineering, and essentially any type of work
that requires calculations to give very precise
solutions. The point of numerical analysis is to
analyse methods that are used to give
approximate number of solutions to situations
where it is unlikely to find the real solution quickly.
 Solution of Linear System of
Equations
• Introduction
• Direct Methods
– LU Decomposition Method
– Crout’s Method
– Choleski’s Method
• Indirect Methods
– Jacobi’s Method
– Gauss-Seidal Method
 Problems involving simultaneous linear
equations often arise in all branches of
sciences and engineering, for example,
electric circuit analysis, vibrations, radiation,
heat transfer and stress analysis in
structure.
In Order to obtain a solution, various
methods have been introduced. This shows
that not a single method exist that could
work in all situations. These methods can be
judge from their speed and accuracy.
 Linear System of Equations
• A linear equation in x1, x2, …,xN is obtained by
requiring the linear combination of the variables to
take on a prescribed value b that is
a1x1 + a2x2 +…+aNxN = b
where ak is the coefficient of xk for k = 1,2,..,N
• System of M linear equations in N unknowns is given
by:
a11x1 + a12x2 +…+a1NxN = b1
a21x1 + a22x2 +…+a2NxN = b2
.
.
aM1x1 + aM2x2 +…+aMNxN = bM
 LU-Decomposition Method
Definition: A square matrix A can be factor into the
product of lower and upper triangular matrices.
That is,
A = LU
where L is the lower triangular matrix and U
represents the upper triangular matrix.
This method is based on the above mentioned
definition. For this reason, it is called as
LU-decomposition method.
As we know that any system of equation can be written in
the form
Ax = b ----------- (1)
Since A = LU, therefore Eq. (1) becomes
LUx = b ----------- (2)
Now set Ux = y, -------------- (3)
Using Eq. (2) in Eq. (3), we get
Ly = b -------------- (4)
where y is the column matrix of order nx1.
Eq. (4) will give the solution by forward substitution and
Eq. (3) will be solved by backward substitution.
*Note: In this method the diagonal entries of the lower
triangular matrix will be kept 1.
 Example 1
Solve the system of equations by LU-
decomposition method
2x1 + x2 +3x3 = 11
4x1 + 3x2 + 10x3 = 28
2x1 + 4x2 + 17x3 = 31
 Crout’s Method
In this method the upper triangular matrix U
has unit diagonal elements whereas the
diagonal elements of L are unrestricted.
Example 2: Solve example 1 by Crout’s
method.
 Choleski’s Method
In this method the corresponding diagonal
elements of L and U are equal.

Example 3: Solve example 1 by Choleski’s

method.
End of Lecture