0% found this document useful (0 votes)

53 views188 pages

Numerical Methods PS

The document outlines the syllabus for a Numerical Methods, Probability and Statistics course at Mohan Babu University, detailing topics such as algebraic equations, numerical differentiation, and statistical significance tests. It includes a list of required textbooks and reference materials, as well as a focus on numerical methods like the Bisection Method for solving equations. The course is scheduled for July to December 2024 and spans various mathematical concepts essential for engineering and scientific applications.

Uploaded by

Subhadip Ghosh

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)

53 views188 pages

Numerical Methods PS

Uploaded by

Subhadip Ghosh

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/ 188

Numerical Methods, Probability and Statistics(NMPS)

Subhadip Ghosh

Department of Mathematics
Mohan Babu University

Subhadip Ghosh (MBU) NMPS July- Dec 2024 1 / 188

Table of Contents

1 Algebraic and Transcendental Equations and Interpolation(10 periods)

2 Numerical Differentiation and Integration(09 periods)

3 Random Variable and Distributions(09 periods)

4 Test of Significance for Large Sample(09 periods)

5 Test of Significance for Small Samples(08 periods)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 2 / 188

Text Books

1 B.S. Grewal, Higher Engineering Mathematics, Khanna Publishers

44/e, 2019
2 S. S. Sastry Introduction Methods of Numerical Analysis, Prentice
Hall of India, 5th edition, 2012
3 Miller and Freund’s , Probability and Statistics for Engineers, Prentice
Hall of India, 8th edition, 2011

Subhadip Ghosh (MBU) NMPS July- Dec 2024 3 / 188

Reference Books

1 S. D. Conte and Carl de Boor, Elementary Numerical Analysis- An

Algorithmic Approach (3rd Edition), McGraw-Hill, 1980
2 . Kreyszig, Advanced engineering mathematics (8th Edition), John
Wiley (1999)
3 C. E. Froberg, Introduction to Numerical Analysis (2nd
Edition),Addison-Wesley, 1981
4 D. Kinciad and W. Cheney, Numerical Analysis and mathematics of
Scientific Computing, Brooks/Cole, 1999.
5 K. E. Atkinson, An Introduction to Numerical Analysis, John-Wiley
and Sons, 2nd Edition, 1989.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 4 / 188

Table of Contents

1 Algebraic and Transcendental Equations and Interpolation(10 periods)

2 Numerical Differentiation and Integration(09 periods)

3 Random Variable and Distributions(09 periods)

4 Test of Significance for Large Sample(09 periods)

5 Test of Significance for Small Samples(08 periods)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 5 / 188

Syllabus for Module-1

1 Bisection Method
2 Newton-Raphson Method
3 Finite Differences
4 Relation Between Operators
5 Newton’s Forward and Backward Difference Formula
6 Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 6 / 188

Solution of Algebraic and Transcendental Equations

A polynomial equation of the form

pn (x) = a0 x n + a1 x n−1 + a2 x n−2 + . . . + an−1 x + an = 0

is called an algebraic equation.

For example:
x 4 − 4x 2 + 5 = 0
4x 2 − 5x + 7 = 0
2x 3 − 5x 2 + 7x + 5 = 0
An equation that contains polynomials, trigonometric functions,
logarithmic functions, exponential functions, etc., is called a
transcendental equation.
For example:
tan x − e x = 0
sin x − xe 2x = 0
xe x = cos x
Subhadip Ghosh (MBU) NMPS July- Dec 2024 7 / 188
Solution of Algebraic and Transcendental Equations

Finding the roots of an equation of the form f (x) = 0 is an important

problem in science and engineering. We assume that f (x) is continuous in
the required interval. A root of an equation f (x) = 0 is the value of x, say
x = α, for which f (α) = 0.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 8 / 188

Solution of Algebraic and Transcendental Equations

A polynomial equation of degree n will have exactly n roots, real or

complex, simple or multiple. A transcendental equation may have one
root, no root, or an infinite number of roots depending on the form of
f (x). The methods of finding the roots of f (x) = 0 are classified as:
1 Direct Methods
2 Numerical Methods
There are no direct methods for solving higher degree algebraic equations
or transcendental equations. Such equations can be solved by Numerical
methods.Here we discuss a few important numerical methods to find a
root of f (x) = 0.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 9 / 188

Bisection Method

This is a very simple method. Identify two points x = a and x = b such

that f (a) and f (b) are having opposite signs. Let f (a) be negative and
f (b) be positive. Then there will be a root of f (x) = 0 in between a and
b. Let the first approximation be the midpoint of the interval (a, b). i.e.
x1 = a+b2

Subhadip Ghosh (MBU) NMPS July- Dec 2024 10 / 188

Bisection Method

If f (x1 ) = 0, then x1 is a root, otherwise the root lies between a and x1 or

x1 and b according to whether f (x1 ) is positive or negative. Then again
we bisect the interval and continue the process until the root is found to
the desired accuracy. Let f (x1 ) be positive, then the root lies in between a
and x1 . The second approximation to the root is given by,
a + x1
x2 =
2
If f (x2 ) is negative, then the next approximation is given by
x2 + x1
x3 =
2
Similarly, we can get other approximations.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 11 / 188

Bisection Method

Example 1: Apply the bisection method to find a root of the equation

x 4 + 2x 3 − x − 1 = 0
Solution: f (x) = x 4 + 2x 3 − x − 1
Here f (0) = −1 and f (1) = 1 ⇒ f (0) · f (1) < 0. Also, f (x) is continuous
in [0, 1], ∴ at least one root exists in [0, 1].
Initial approximation: a = 0, b = 1
x0 = 0+12 = 0.5, f (0.5) = −1.1875, f (0.5) · f (1) < 0
First approximation: a = 0.5, b = 1
x1 = 0.5+1
2 = 0.75, f (0.75) = −0.5898, f (0.75) · f (1) < 0
Second approximation: a = 0.75, b = 1
x2 = 0.75+1
2 = 0.875, f (0.875) = 0.051, f (0.75) · f (0.875) < 0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 12 / 188

Bisection Method

Third approximation: a = 0.75, b = 0.875

x3 = 0.75+0.875
2 = 0.8125, f (0.8125) = −0.30394, f (0.8125) · f (0.875) < 0
Fourth approximation: a = 0.8125, b = 0.875
x4 = 0.8125+0.875
2 = 0.84375, f (0.84375) = −0.135,
f (0.84375) · f (0.875) < 0
Fifth approximation: a = 0.84375, b = 0.875
x5 = 0.84375+0.875
2 = 0.8594, f (0.8594) = −0.0445,
f (0.8594) · f (0.875) < 0
Sixth approximation: a = 0.8594, b = 0.875
x6 = 0.8594+0.875
2 = 0.8672, f (0.8672) = 0.0027, f (0.8594) · f (0.8672) < 0
Seventh approximation: a = 0.8594, b = 0.8672
x7 = 0.8594+0.8672
2 = 0.8633
The first two decimal places have been stabilized; hence 0.8633 is the real
root correct to two decimal places.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 13 / 188

Bisection Method

Example 2
Find a real root of the equation f (x) = x 3 − x − 1 = 0, using the
Bisection method.
Solution:
First find the interval in which the root lies, by trial and error method.
f (1) = 13 − 1 − 1 = −1, which is negative
f (2) = 23 − 2 − 1 = 5, which is positive
∴ A root of f (x) = x 3 − x − 1 = 0 lies in between 1 and 2.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 14 / 188

Bisection Method

1+2
∴ x1 = = 1.5
2
f (x1 ) = f (1.5) = (1.5)3 − 1.5 − 1 = 0.875, which is positive.
Hence, the root lies in between 1 and 1.5.
1 + 1.5
∴ x2 = = 1.25
2
f (x2 ) = f (1.25) = (1.25)3 − 1.25 − 1 = −0.29, which is negative.
Hence, the root lies in between 1.25 and 1.5.
1.25 + 1.5
∴ x3 = = 1.375
2
Similarly, we get x4 = 1.3125, x5 = 1.34375, x6 = 1.328125, etc.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 15 / 188

Bisection Method

Example 3
Find a root of f (x) = xe x − 1 = 0, using the Bisection method, correct to
three decimal places.
Solution:
f (0) = 0 · e 0 − 1 = −1 < 0
f (1) = 1 · e 1 − 1 = 1.7183 > 0
Hence, a root of f (x) = 0 lies in between 0 and 1.
∴ x1 = 0+12 = 0.5
f (0.5) = 0.5 · e 0.5 − 1 = −0.1756
Hence, the root lies between 0.5 and 1.
∴ x2 = 0.5+1
2 = 0.75

Subhadip Ghosh (MBU) NMPS July- Dec 2024 16 / 188

Bisection Method

Proceeding like this, we get the sequence of approximations as follows:

x3 = 0.625, x4 = 0.5625, x5 = 0.59375, x6 = 0.5781, x7 = 0.5703,
x8 = 0.5664, x9 = 0.5684, x10 = 0.5674, x11 = 0.5669, x12 = 0.5672,
x13 = 0.5671
Hence, the required root correct to three decimal places is x = 0.567.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 17 / 188

Bisection Method

Example 4
Using the bisection method, find an approximate root of the equation
sin x = x1 correct to two decimal places.
Solution: f (x) = x sin x − 1
Here f (1) = sin 1 − 1 = −0.1585 and f (2) = 2 sin 2 − 1 = 0.8186
Also f (x) is continuous in [1, 2], ∴ at least one root exists in [1, 2].
Initial approximation: a = 1, b = 2
x0 = 1+22 = 1.5, f (1.5) = 0.4963, f (1) · f (1.5) < 0
First approximation: a = 1, b = 1.5
x1 = 1+1.5
2 = 1.25, f (1.25) = 0.1862, f (1) · f (1.25) < 0
Second approximation: a = 1, b = 1.25
x2 = 1+1.25
2 = 1.125, f (1.125) = 0.0151, f (1) · f (1.125) < 0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 18 / 188

Bisection Method

Third approximation: a = 1, b = 1.125

x3 = 1+1.125
2 = 1.0625, f (1.0625) = −0.0718, f (1.0625) · f (1.125) < 0
Fourth approximation: a = 1.0625, b = 1.125
x4 = 1.0625+1.125
2 = 1.09375, f (1.09375) = −0.0284,
f (1.09375) · f (1.125) < 0
Fifth approximation: a = 1.09375, b = 1.125
x5 = 1.09375+1.125
2 = 1.10937, f (1.10937) = −0.0066,
f (1.10937) · f (1.125) < 0
Sixth approximation: a = 1.10937, b = 1.125
x6 = 1.10937+1.125
2 = 1.11719, f (1.11719) = 0.0042,
f (1.10937) · f (1.11719) < 0
Seventh approximation: a = 1.10937, b = 1.11719
x7 = 1.10937+1.11719
2 = 1.11328, f (1.11328) = −0.0012 ≈ 0
Hence 1.11328 is the real root correct to two decimal places.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 19 / 188

Bisection Method

Example 5
Apply the bisection method to find a root of the equation
x 3 − 2x 2 − 4 = 0 correct to three decimal places.
Solution: f (x) = x 3 − 2x 2 − 4 Here f (2) = −4 and
f (3) = 5 ⇒ f (2) · f (3) < 0. Also, f (x) is continuous in [2, 3], ∴ at least
one root exists in [2, 3].
Initial approximation: a = 2, b = 3
x0 = 2+32 = 2.5, f (2.5) = −1.8750, f (2.5) · f (3) < 0
First approximation: a = 2.5, b = 3
x1 = 2.5+3
2 = 2.75, f (2.75) = 1.6719, f (2.5) · f (2.75) < 0
Second approximation: a = 2.5, b = 2.75
x2 = 2.5+2.75
2 = 2.625, f (2.625) = 0.3066, f (2.5) · f (2.625) < 0
Third approximation: a = 2.5, b = 2.625
x3 = 2.5+2.625
2 = 2.5625, f (2.5625) = −0.3640, f (2.5625) · f (2.625) < 0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 20 / 188

Bisection Method

Fourth approximation: a = 2.5625, b = 2.625

x4 = 2.5625+2.625
2 = 2.59375, f (2.59375) = −0.0055,
f (2.59375) · f (2.625) < 0
Fifth approximation: a = 2.59375, b = 2.625
x5 = 2.59375+2.625
2 = 2.60938, f (2.60938) = 0.1488,
f (2.59375) · f (2.60938) < 0
Sixth approximation: a = 2.59375, b = 2.60938
x6 = 2.59375+2.60938
2 = 2.60157, f (2.60157) = 0.0719,
f (2.59375) · f (2.60157) < 0
Seventh approximation: a = 2.59375, b = 2.60157
x7 = 2.59375+2.60157
2 = 2.59765, f (2.59765) = 0.0329,
f (2.59375) · f (2.59765) < 0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 21 / 188

Bisection Method

Eighth approximation: a = 2.59375, b = 2.59765

x8 = 2.59375+2.59765
2 = 2.5957, f (2.5957) = 0.0136,
f (2.59375) · f (2.5957) < 0
Ninth approximation: a = 2.59375, b = 2.5957
x9 = 2.59375+2.5957
2 = 2.5947, f (2.5947) = −0.004,
f (2.5947) · f (2.5957) < 0
Tenth approximation: a = 2.5947, b = 2.5957
x10 = 2.5947+2.5957
2 = 2.5952
Hence, 2.5952 is the real root correct to three decimal places.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 22 / 188

Drawback of Bisection Method

1 The convergence of the bisection method is slow as it is simply based

on halving the interval.
2 If one of the initial guesses is closer to the root, it will take a larger
number of iterations to reach the root.
3 If a function f (x) is such that it just touches the x-axis (see Figure
1), such as f (x) = x 2 , it will be unable to find the lower guess, xℓ ,
and upper guess, xu , such that f (xℓ ) · f (xu ) < 0.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 23 / 188

Bisection Method

Figure: Graph of f (x) = x 2 touching the x-axis

Subhadip Ghosh (MBU) NMPS July- Dec 2024 24 / 188

Newton-Raphson’s Method

This is another important method. Let x0 be an approximation for the

root of f (x) = 0. Let x1 = x0 + h be the correct root such that f (x1 ) = 0.
Expanding f (x1 ) = f (x0 + h) by Taylor series, we get:

h2
f (x1 ) = f (x0 + h) = f (x0 ) + hf ′ (x0 ) + + ... = 0 (1)
2!
For small values of h, neglecting the terms with h2 , h3 , etc.,
we get:
f (x0 ) + hf ′ (x0 ) = 0 (2)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 25 / 188

Newton-Raphson’s Method

f (x)
h=−
f ′ (x)
f (x)
∴ x1 = x0 + h = x0 −
f ′ (x)
Proceeding like this, successive approximations x2 , x3 , . . . , xn+1 are given
by:
f (xn )
xn+1 = xn − ′ (3)
f (xn )
For n = 0, 1, 2, . . .

Subhadip Ghosh (MBU) NMPS July- Dec 2024 26 / 188

Newton-Raphson’s Method

EXAMPLE 1
Find a root of the equation x 2 − 2x − 5 = 0 by Newton-Raphson method.
SOLUTION:
Here f (x) = x 3 − 2x − 5. Therefore, f ′ (x) = 3x 2 − 2.
Newton-Raphson method formula is

f (xn )
xn+1 = xn −
f ′ (xn )

where f (x) = x 3 − 2x − 5.

3xn3 − 2xn − 5
∴ xn+1 = xn − , n = 0, 1, 2, . . . (1)
3xn2 − 2

Subhadip Ghosh (MBU) NMPS July- Dec 2024 27 / 188

Newton-Raphson’s Method

Let x0 = 2.
Therefore, f (x0 ) = f (2) = 23 − 2 · 2 − 5 = −1
and f ′ (x0 ) = f ′ (2) = 3 · 22 − 2 = 10.
Putting n = 0 in (1), we get

−1
x1 = 2 − = 2.1
10
f (x1 ) = f (2.1) = 2.13 − 2 · 2.1 − 5 = 0.061
f ′ (x1 ) = f ′ (2.1) = 3 · 2.12 − 2 = 11.23
Therefore,
0.061
x2 = 2.1 − = 2.094568
11.23
Similarly, we can calculate x3 , x4 , . . .

Subhadip Ghosh (MBU) NMPS July- Dec 2024 28 / 188

Newton-Raphson’s Method

Example 2 Find a root of x sin x + cos x = 0 using the Newton-Raphson

method.
Solution:
f (x) = x sin x + cos x.
∴ f ′ (x) = sin x + x cos x − sin x = x cos x
The Newton-Raphson method formula is,
xn+1 = xn − xn sin xn +cos xn
xn cos xn , n = 0, 1, 2, . . .
Let x0 = π = 3.1416.
∴ x1 = 3.1416 − 3.1416 sin π+cos π
3.1416 cos π = 2.8233.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 29 / 188

Newton-Raphson’s Method

Similarly,
x2 = 2.7986
x3 = 2.7984
x4 = 2.7984
∴ x = 2.7984 can be taken as a root of the equation x sin x + cos x = 0.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 30 / 188

Newton-Raphson’s Method

Example 3
Find the smallest positive root of x − e −x = 0, using the Newton-Raphson
method.
Solution
Here,
f (x) = x − e −x
f ′ (x) = 1 + e −x
f (0) = −1 and f (1) = 0.63212.
∴ The smallest positive root of f (x) = 0 lies between 0 and 1. Let x0 = 1
The Newton-Raphson method formula is,
−xn
n −e
xn+1 = xn − x1+e −xn , n = 0, 1, 2, . . .

Subhadip Ghosh (MBU) NMPS July- Dec 2024 31 / 188

Newton-Raphson’s Method

f (0) = f (1) = 0.63212,

f ′ (0) = f ′ (1) = 1.3679
∴ x1 = x0 − ff ′(x 0) 0.63212
(x0 ) = 1 − 1.3679 = 0.5379. f (0.5379) = −0.0461,
f ′ (0.5379) = 1.584
∴ x2 = 0.5379 + 0.0461 1.584 = 0.567.
Similarly, x3 = 0.56714.
∴ x = 0.567 can be taken as the smallest positive root of x − e −x = 0,
correct to three decimal places.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 32 / 188

Newton-Raphson’s Method

Example 4
Using the Newton-Raphson method: (a) Find the square root of a
number. (b) Find the reciprocal of a number. [JNTU 2008]
Solution:
√
(a) Let n be the number and x = n ⇒ x 2 = n.
If f (x) = x 2 − n = 0, . . . (1)
√
Then the solution to f (x) = x 2 − n = 0 is x = n.
f ′ (x) = 2x
By the Newton-Raphson method:
xi2 −n

xi+1 = xi − ff ′(xi)
(xi ) = xi − 2x i
= 1
2 xi + n
xi
. . . (2)
Using the above formula the square root of any number ‘n’ can be found
to required accuracy

Subhadip Ghosh (MBU) NMPS July- Dec 2024 33 / 188

Newton-Raphson’s Method

1
(b) To find the reciprocal of a number n, let f (x) = x − n = 0. . . . (1)
∴ The solution of (1) is x = n1 .
f ′ (x) = − x12
Now, by the Newton-Raphson method:
1
f (xi ) xi
−n
xi+1 = xi − f ′ (xi ) = xi − − 1 = xi (2 − xi n)
x2
i
Using the above formula the reciprocal of a number can be found to
required accuracy.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 34 / 188

Newton-Raphson’s Method

Example 5
Find a real root for x tan x +1 = 0 using Newton–Raphson method
[JNTU 2006]
Solution:
Given f (x) = x tan x + 1 = 0
f ′ (x) = x sec2 x + tan x
f (2) = 2 tan 2 + 1 = −3.370079 < 0
f (3) = 2 tan 3 + 1 = −0.572370 > 0
∴ The root lies between 2 and 3.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 35 / 188

Newton-Raphson’s Method

Take x0 = 2+3 2 = 2.5 (average of 2 and 3).

By the Newton-Raphson method:
x1 = x0 − ff ′(x0) 0.86755
(x0 ) = 2.5 − 3.14808 = 2.77558

Subhadip Ghosh (MBU) NMPS July- Dec 2024 36 / 188

Exercises

1 Using the Bisection method, find the smallest positive root of

x 3 − x − 4 = 0 correct to two decimal places. [Ans: 1.80]
2 Obtain a root correct to three decimal places of x 3 − 18 = 0, using
the Bisection Method. [Ans: 2.621]
3 Find a root of the equation xe x − 1 = 0 which lies in (0, 1), using the
Bisection Method. [Ans: 0.567]
4 Find a root of x 4 − x − 10 = 0 by the Newton-Raphson Method.
[Ans: 1.856]
5 Find a real root of x − cos x = 0 by the Newton-Raphson Method.
[Ans: 0.739]
6 Find a root of 2x − 3 sin x − 5 = 0 by the Newton-Raphson Method.
[Ans: 2.883238]
7 Find the smallest positive root of tan x = x by the Newton-Raphson
Method. [Ans: 4.4934]

Subhadip Ghosh (MBU) NMPS July- Dec 2024 37 / 188

Finite Differences

Let y = f (x) be a real valued function of x defined in an interval [a, b] and

its values are known for (n + 1) equally spacing points xi (i = 0, 1, 2, ...)
such that xi = x0 + ih(i = 0, 1, 2, ..., n) where x0 = a, xn = b and h is
space length. Then xi (i = 0, 1, 2, ...) are called nodes and the
corresponding values yi are termed as entries.
Now we introduce the concept of various type differences in order to find
the values of f (x) or its derivative for some intermediate values of
x ∈ [a, b]

Subhadip Ghosh (MBU) NMPS July- Dec 2024 38 / 188

Forward Differences

The differences y1 − y0 , y2 − y1 , y3 − y2 , . . . , yn − yn−1 for the entries

y0 , y1 , y2 , . . . , yn−1 , yn are called the first forward differences and are
denoted by ∆y0 , ∆y1 , ∆y2 , . . . , ∆yn−1 respectively. Thus we have
∆yi = yi+1 − yi (i = 0, 1, 2, . . . , n − 1)
where ∆ is called the forward difference operator. In general, the forward
difference operator is defined by
∆f (x) = f (x + h) − f (x)
Similarly, the higher order forward differences are defined as
∆2 yi = ∆yi+1 − ∆yi

Subhadip Ghosh (MBU) NMPS July- Dec 2024 39 / 188

Forward Differences

The third-order forward differences are defined as:

∆3 yi = ∆2 yi+1 − ∆2 yi
In general, the r -th order forward difference is given by:
∆r yi = ∆r −1 yi+1 − ∆r −1 yi
where i = 0, 1, 2, . . . , n − 1 and r (1 ≤ r ≤ n) is a positive integer.
Now, the second-order forward difference for y0 is:
∆2 y0 = ∆y1 − ∆y0 = (y2 − y1 ) − (y1 − y0 ) = y2 − 2y1 + y0
The third-order forward difference for y0 is:
∆3 y0 = ∆2 y1 −∆2 y0 = (y3 −2y2 +y1 )−(y2 −2y1 +y0 ) = y3 −3y2 +3y1 −y0
Hence, for the n-th
order forward difference, we have:
∆n y0 = yn − n1 yn−1 + n2 yn−2 − · · · + (−1)n y0

We can calculate the above forward differences easily with the help of the
following tables, called the forward difference table.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 40 / 188

Forward Differences

Subhadip Ghosh (MBU) NMPS July- Dec 2024 41 / 188

Forward Differences

Let a and b be any two constants. Then, the following properties hold:
1 Property 1: ∆a = 0
2 Property 2: ∆{af (x)} = a∆f (x)
3 Property 3: ∆{af (x) ± bg (x)} = a∆f (x) ± b∆g (x)
4 Property 4: ∆[f (x)g (x)] = f (x)∆g (x) + ∆f (x).g (x + h) =
f (x + h)∆g (x) + ∆f (x).g (x)
5 Property 5: ∆ gf (x)
(x) =
g (x)∆f (x)−f (x)∆g (x)
g (x)g (x+h)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 42 / 188

Backward Differences

The differences y1 − y0 , y2 − y1 , y3 − y2 , . . . yn , yn−1 for the entries

y1 , y2 , y3 , . . . , yn are called the first backward differences and are denoted
by ∇y1 , ∇y2 , ∇y3 , . . . respectively.
Thus we have
∇yi = yi − yi−1 , i = 1, 2, . . . , n
where ∇ is called the backward difference operator. In general, the
backward difference operator is defined by
∇f (x) = f (x) − f (x − h)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 43 / 188

Backward Differences

Similarly, the higher order backward differences are defined as:

∇2 yi = ∇yi − ∇yi−1
∇3 yi = ∇2 yi − ∇2 yi−1
and so on.
Now, let us compute some specific differences:
∇y1 : ∇y1 = y1 − y0
For ∇2 y2 : ∇2 y2 = ∇y2 − ∇y1 = (y2 − y1 ) − (y1 − y0 )
For ∇3 y3 : ∇3 y3 = y3 − 3y2 + 3y1 − y0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 44 / 188

Backward Differences

For the nth order backward difference, we have:

n n
∇ yn = yn − 1 yn−1 + 2 yn−2 + · · · + (−1)n y0
n

where kn represents the binomial coefficient.

We can calculate the different order backward differences very quickly with
the help of the following table, called the backward difference table:

Subhadip Ghosh (MBU) NMPS July- Dec 2024 45 / 188

Relation between operators

Forward difference operator: ∆f (x) = f (x + h) − f (x)

Backward difference operator: ∇f (x) = f (x) − f (x − h)
Show that ∆.∇ ≡ ∆ − ∇

∆.∇f (x) = ∆[f (x) − f (x − h)]

= ∆f (x) − ∆f (x − h)
= ∆f (x) − [f (x) − f (x − h)]
= ∆f (x) − ∇f (x)
= (∆ − ∇)f (x)

we have shown that: ∆.∇ ≡ ∆ − ∇

Subhadip Ghosh (MBU) NMPS July- Dec 2024 46 / 188

Relation between operators

The Shift operator is denoted by E and is defined as follows:

Ef (x) = f (x + h), where h is the spacing.
E 2 f (x) = E [f (x + h)] = f (x + 2h)
E 3 f (x) = E [f (x + 2h)] = f (x + 3h)
In general, we have:

E n f (x) = f (x + nh)
The inverse shift operator E −1 is defined by:
E −1 f (x) = f (x − h)
In general, E −n f (x) = f (x − nh)

Subhadip Ghosh (MBU) NMPS July- Dec 2024 47 / 188

Relation between operators

Since the forward difference operator ∆ is defined as:

∆f (x) = f (x + h) − f (x)
it follows that:

Ef (x) = f (x + h) = ∆f (x) + f (x)

Rearranging gives:

Ef (x) = (∆ + 1)f (x)

Thus:

∆=E −1

Subhadip Ghosh (MBU) NMPS July- Dec 2024 48 / 188

Relation between operators

We know that:

∆2 f (x) = ∆[∆f (x)] = ∆[f (x + h) − f (x)] = f (x + 2h) − 2f (x + h) + f (x)

Using the shift operator E , we can express this as:

∆2 f (x) = E 2 f (x) − 2Ef (x) + f (x) = (E − 1)2 f (x)

Thus:

∆2 ≡ (E − 1)2
In general, we have:

∆n ≡ (E − 1)n

Subhadip Ghosh (MBU) NMPS July- Dec 2024 49 / 188

Frame Title

Additionally, for the inverse shift operator E −1 :

E −1 f (x) = f (x − h) = f (x) − ∇f (x)

Rearranging gives:

∇f (x) = f (x) − f (x − h)
Thus,
E −1 f (x) = f (x) − ∇f (x) = (1 − ∇)f (x)
And

E −1 = 1 − ∇

Subhadip Ghosh (MBU) NMPS July- Dec 2024 50 / 188

Interpolation Using Newton Forward and Backward
Difference Formulae

The Newton-Gregory formula can be written as follows:

f (x + nh) = E n f (x) (1)

n
= (1 + ∆) f (x) (2)
n
X n
= ∆i f (x) (3)
i
i=0

n n 2 n
= f (x) + ∆f (x) + ∆ f (x) + · · · + ∆n f (x) (4)
1 2 n
n
X n
= ∆i f (x) (5)
i
i=0

Subhadip Ghosh (MBU) NMPS July- Dec 2024 51 / 188

Interpolation Using Newton Forward Difference Formulae

Ex-1
If y (0) = −1, y (1) = 3, y (2) = 8, y (3) = 13 find y(6).
Solution:
Let us construct the following difference table;

x y ∆y ∆2 y ∆3 y
0 −1
4
1 3 1
5 −1
2 8 0
5
3 13

Subhadip Ghosh (MBU) NMPS July- Dec 2024 52 / 188

Interpolation Using Newton Forward and Backward
Difference Formulae

Using the Newton-Gregory formula, we have:

f (x + nh) = ni=0 ni ∆i f (x)
P

To find f (6), we use xP = 0, h = 1, and n = 6:

f (6) = f (0 + 6 · 1) = 6i=0 6i ∆i f (0)
Calculating each term:
f (0) = −1
∆f (0) = 4 (from the table)
∆2 f (0) = 1
∆3 f (0) = −1

Subhadip Ghosh (MBU) NMPS July- Dec 2024 53 / 188

Interpolation Using Newton Forward and Backward
Difference Formulae

Therefore:
f (6) = f (0) + 61 ∆f (0) + 62 ∆2 f (0) + 6
∆3 f (0)

3
f (6) = −1 + 6 · 4 + 15 · 1 + 20 · (−1)
f (6) = −1 + 24 + 15 − 20 = 18
Thus, f (6) = 18.

Subhadip Ghosh (MBU) NMPS July- Dec 2024 54 / 188

Central Difference and Average Operator

The central difference operator is denoted by δ and is defined as:

h h
δf (x) = f x + −f x −
2 2
We can write this as:

δf (x) = E 1/2 f (x) − E −1/2 f (x) = E 1/2 − E −1/2 f (x)

where
δ = E 1/2 − E −1/2

Subhadip Ghosh (MBU) NMPS July- Dec 2024 55 / 188

Central Difference and Average Operator

The average operator µ is defined as:

1 h h
µf (x) = f x+ +f x −
2 2 2
Using the shift operators, this can be expressed as:
1 1/2
µf (x) =
E + E −1/2 f (x)
2
Thus, the average operator µ can be written as:
1 1/2
µ= E + E −1/2
2

Subhadip Ghosh (MBU) NMPS July- Dec 2024 56 / 188

Interpolation with unequal interval:Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 57 / 188

Interpolation with unequal interval:Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 58 / 188

Interpolation with unequal interval:Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 59 / 188

Interpolation with unequal interval:Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 60 / 188

Interpolation with unequal interval:Lagrange’s Formula

Subhadip Ghosh (MBU) NMPS July- Dec 2024 61 / 188

Table of Contents

1 Algebraic and Transcendental Equations and Interpolation(10 periods)

2 Numerical Differentiation and Integration(09 periods)