Chapter 8

Numerical Differentiation
&
Integration
8.1 Introduction
Differentiation and integration are basic mathematical operations with a wide
range of applications in various fields of science and engineering. Simple
continuous algebraic or transcendental functions can be easily differentiated or
integrated directly. However at times there are complicated continuous functions
which are tedious to differentiate or integrate directly or in the case of
experimental data, where tabulated values of variables are given in discrete form,
direct methods of calculus are not applicable.
In this chapter, we develop ways to approximate the derivatives of function
, when only data points are given and also to integrate definite integrals
by splitting the area under the curve in specified ways.
8.2 Numerical Differentiation
Numerical differentiation is the process of computing the value of the derivative of
an explicitly unknown function, with given discrete set of points
. To differentiate a function numerically, we first determine an
interpolating polynomial and then compute the approximate derivative at the given
point.
If ’s are equispaced
i. Newton's forward interpolation formula is used to find the derivative near the
beginning of the table.
ii. Newton's backward interpolation formula is used to compute the derivation
near the end of the table.
iii. Stirling’s formula is used to estimate the derivative near the centre of the table.
If ’s are not equispaced, we may find using Newton’s divided difference
method or Lagrange’s interpolation formula and then differentiate it as many times
as required.
8.2.1 Derivatives Using Newton’s Forward Interpolation Formula
Newton’s forward interpolation formula for the function is given by

,
Differentiating with respect to

Also

Again

Using , we get

Now at ,

Now

Or

8.2.2 Derivatives Using Newton’s Backward Interpolation Formula

Newton’s backward interpolation formula for the function is given by

Differentiating with respect to

Also

Again

Using , we get

Now at ,
Similarly

8.2.3 Derivatives Using Stirling’s Interpolation Formula

Stirling’s central difference interpolation formula (taking as the middle value
of the table) is given by:

Differentiating with respect to

Also

Again

Using , we get

Now at ,

Similarly

8.2.4 Derivatives of Polynomial with Unequispaced ’s

In case where ’s are not equispaced in a given data, the polynomial
may be found using Newton’s divided difference method or Lagrange’s
interpolation formula and then direct differentiation may be applied.
Remark:

 In case derivative has to be found at a point, whose value needs to be

interpolated, then first apply applicable interpolation technique and then
differentiate the function.

 In all the cases irrespective of data points being equispaced or not, the
polynomial may be found using the applicable interpolation
formulae and then direct differentiation can be done using usual calculus
techniques.
Example 1 Given a cubic polynomial with following data points
0 1 2 3
5 6 3 8

Find and at

Solution: Derivative has to be evaluated near the starting of the table, thereby
constructing forward difference table for the function

To find the derivative at , taking and applying the relation:

From table , , , ,
Substituting these values in , we get

Also

Aliter
Newton’s forward interpolation formula given by:

, ,
Also from table , ,
Substituting these values in , we get
 Also

Example 2 Given a polynomial with following data points:

1.0 1.1 1.2 1.3 1.4 1.5 1.6
7.989 8.403 8.781 9.129 9.451 9.750 10.031

Find and at and

Solution: Derivatives has to be evaluated near the starting as well as towards the
end of the table, thereby constructing difference table for the function

7.989

8.403
0.378

0.002

To find the derivative at , taking and applying the relation:

From table , , , ,
,
Substituting these values in , we get

Also

To find the derivative at , taking and applying the relation:

From table , , , ,
,
Substituting these values in , we get

Also

Example3 Following table gives the census population of a state for the years
to .
Year 1961 1971 1981 1991 2001
Population
19.96 36.65 58.81 77.21 94.61
(Million)
Find the rate of growth of the population in the year 2001
Solution: Derivative has to be evaluated near the end of the table, thereby
constructing backward difference table for the function
Year Population

1971

To find rate of growth (derivative) in the year 2001, taking and

applying the relation:

From table , , , ,
Substituting these values in , we get
Example4 A slider in a machine moves along a fixed straight rod. Its
distance cm along the rod is given below for various values of the time . Find the
velocity and acceleration of the slider when seconds.
0 0.1 0.2 0.3 0.4 0.5 0.6
30.13 31.62 32.87 33.64 33.95 33.81 33.24
Solution: Derivatives has to be evaluated towards the centre of the table, thereby
constructing central difference table for the function . Also , ,
, 3 , 3, 4, 5, 6 lie along the dotted line as shown.

30.13

31.62
1.25

To find the velocity (derivative) at , taking and applying the

relation:

From table , , , , ,
., All the positions have been shown, enclosed in
boxes.
Substituting these values in , we get
seconds

To find the acceleration (second derivative) at and applying the relation

Also from table, ,, ,

Substituting these values in , we get
seconds2
Example5 The following data gives corresponding values of pressure ‘ ’ and
specific volume ‘ ’ of a super heated expandable system
2 4 6 8 10
105 42.7 25.3 16.7 13
Find the rate of change of volume with respect to pressure, when units.
Solution: Values of are not equispaced, thereby constructing divided difference
table for the function

105

42.7 0.001

25.3 0.005 0.000007

16.7 0.025

Newton’s divided difference formula is given by

Here , , , ,
, , ,

Substituting these values in , we get

25.3(1)](0.000007)

It is evident that volume decreases with increase in pressure.

Example6 Find for the following the following data points of a polynomial
.
0 2 4 6 8
4 8 15 7 6
Solution: To find or , constructing forward difference table for the
function .

0
4
2 3
7
4 40

6 7

Newton’s forward interpolation formula given by:

Also

Again

Using , we get

Now for , , ,
Also from table , , ,
Substituting these values in , we get

Note: Here formula for computing derivatives can not be applied directly as
at point at which derivative has to be computed does not exist in
the table and has to be interpolated first.
8.2.5 Maxima and Minima of a Tabulated Function:
Newton’s forward interpolation formula for the function is given by
 ,

Differentiating with respect to

For finding maxima/minima of a function ,

,

Neglecting 4th and higher order differences in equation and substituting in ,

we get a quadratic equation of the form ,where , , are
constants. Solving for and substituting in , we get points of
maxima/minima for the function .
 Newton’s forward method is apt for finding extreme values of a tabulated
data, wherever their location may be, by index assuming values ,
if the extreme value is not in vicinity of the point ( ). Yet we may also
use Newton’s backward or Stirling’s central differences formulae to locate
extreme values, if desired.
Example 7 From the following data, find maximum and minimum values of .
0 2 4 6
2 0
Solution: Constructing forward difference table for the function

Newton’s forward interpolation formula for the function is given by

Taking , , , ,
Substituting these values in , we get
Differentiating with respect to , we get

For to be maximum,

Substituting in , maximum and minimum values of are 4 and 0 respectively.

Example 8 From the following table, find for which is maximum.
3 4 5 6 7 8
0.205 0.240 0.259 0.262 0.250 0.224
Also find maximum value of .
Solution: Constructing forward difference table for the function , upto
third differences

0.205

0.240
0.019 0

Newton’s forward interpolation formula for the function is given by

Taking , , , ,
Substituting these values in , we get

Differentiating with respect to , we get

For to be maximum,
Also or

is maximum when or
Substituting in , maximum value of is given by

8.3 Numerical Integration

Numerical Integration is the process of computing the value of definite
integral , when the integrand function is given as discrete set of
points . As in case of numerical differentiation, here
also the integrand is first replaced with an interpolating polynomial, and
then it is integrated to compute the value of the definite integral. This gives us
'quadrature formula' for numerical integration.
Newton Cotes Quadrature Formula
For an explicitly known function , let take values , , ,
for taking values , , , , where ’s are equispaced.
To evaluate , divide the interval into equal parts, each of
width , such that
Clearly Then ,

By Newton’s forward interpolation formula:

, also when , , ,

This is known as Newton’s Cote’s quadrature formula. Different quadrature

formulae are derived by taking , , , in equation .
8.3.1Numerical Integration Using Trapezoidal Rule
If we put in Newton’s Cote’s quadrature formula given by , we get the
curve through and as a straight line and being linear equation in ,
nd
2 and higher order differences are zero.

Similarly

Adding areas of all these intervals, we get:

This is known as trapezoidal rule to evaluate , where the function

is given as discrete set of points .
Geometrical Significance of Trapezoidal Rule
In trapezoidal rule, the curve is
replaced by piecewise straight lines joining
the points and ;
and ; ; and .
The area under the curve , between
the ordinates ; and above
axis is approximately equal to the sum of areas
of trapezoids obtained within the enclosed
region, shown by shaded portion of adjoining
figure.
Error in Trapezoidal Rule
Area of first trapezoid in the interval is given by

Also expanding by Taylor’s series about

Putting , we get
Substituting in , we get

Again taking area of single strip, under the curve , between the
ordinates and and above
using

Error in the interval is given by

Thus lowest order error in interval

Similarly lowest error in interval

Lowest error in interval

Total error
If

Then

, where
Hence the error in trapezoidal rule is of order , where is the height of the
interval.
8.3.2Numerical Integration Using Simpson’s One-Third Rule
If we put in Newton’s Cote’s quadrature formula given by , we get the
curve through the points , , as a parabolic figure and
rd
being quadratic equation in , 3 and higher order differences are zero.
 Similarly

Adding areas of all these intervals, we get:

This is known as Simpson’s one-third rule to evaluate , where the

function is given as discrete set of points .
Geometrical Significance of Simpson’s One-Third Rule
In Simpson’s one-third rule, the curve is
nd
replaced by arcs of 2 degree parabolas with vertical
axis as shown in given figure.
Simpson’s one-third rule requires the given interval to
be divided into even number of sub-intervals, since we
are finding areas of two strips at a time.
Error in Simpson’s One-Third Rule
Area of first two strips in the interval is given by

Also expanding by Taylor’s series about

Putting in , we get

Putting in , we get
Substituting and in , may be written as

Again taking actual area of two strip, under the curve , between the
ordinates and and above
using

Error in the interval is given by

Thus lowest order error in interval

Similarly lowest order error in interval

Thus lowest order error in interval

Total error

If

Then number of intervals is

, where
Hence the error in Simpson’s one-third rule is of order , where is the
height of the interval.
8.3.3 Numerical Integration Using Simpson’s Three- Eighth Rule
If we put in Newton’s Cote’s quadrature formula given by , we get the
curve through the points , , , as a cubic
th
polynomial and hence 4 and higher order differences are zero.
Newton’s Cote’s quadrature formula reduces to:

Similarly

Adding areas of all these intervals, we get:

This is known as Simpson’s three-eighths rule to evaluate , where the

function is given as discrete set of points .
Geometrical Significance of Simpson’s Three-Eighth Rule
The Simpson’s 3/8 rule is similar to the 1/3 rule
except that curve is replaced by arcs of
rd
3 degree polynomial curve, as shown in given
figure. It is used when it is required to take 3
segments at a time. Thus number of intervals
must be a multiple of 3.
It can be derived that the lowest order error in interval

number of intervals is

, where
Hence error in Simpson’s 3/8 rule is of order , where is the height of the
interval.
8.3.4 Applications of Numerical Integration
Numerical integration has numerous practical applications in the field of calculus.
Simpson’s rule due to its ease in application and higher accuracy is a preferred
method in various application areas as given below:
Area bounded by a curve between the ordinates and ,
above is given by .
Volume of solid formed by revolving the curve between the
ordinates and along is given by .
Length of an arc of the curve between the ordinate an

and is given b
To find velocity when acceleration at different times is given in tabular form.
To find displacement when velocity is given as a function of time in discrete
form.
Remarks:
 Simpson’s rules ideally returns more accurate results compared to
trapezoidal rule provided is small, less than one essentially.
 Simpson’s rule requires odd number of points (even number of sub-
intervals) for application.
 Simpson’s rule requires number of sub-intervals to be multiple of 3.

Example9 Evaluate using

i. Trapezoidal rule taking

ii. Simpson’s rule taking
iii. Simpson’s rule taking

Solution: To solve using trapezoidal rule

Taking ,
Dividing the interval into 5 equal parts for the function
 0 0.2 0.4 0.6 0.8 1
1 0.96 0.74 0.61 0.5
By trapezoidal rule

using trapezoidal rule.

To solve using Simpson’s rule

Taking ,
Dividing the interval into 4 equal parts for the function

0 0.25 0.5 0.75 1

1 0.94 0.64 0.5
By Simpson’s rule

using Simpson’s rule

To solve using Simpson’s rule

Taking ,

Dividing the interval into 6 equal parts for the function

0 1
1 0.97 0.8 0.69 0.59 0.5
By Simpson’s rule

using Simpson’s rule

Example10 Evaluate using

Trapezoidal rule Simpson’s rule Simpson’s rule, taking

Solution: Taking ,
Dividing the interval into 6 equal parts for the function

0
 0 0.2588 0.7071 0.866 0.9659 1

To solve using trapezoidal rule

using trapezoidal rule.

To solve using Simpson’s rule

using Simpson’s rule

To solve using Simpson’s rule

using Simpson’s rule

Example11 Evaluate using

Trapezoidal rule Simpson’s rule Simpson’s rule, taking

Solution: Taking ,
Dividing the interval into 8 equal parts for

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

3.26 2.04 1.55 1.29 1.12 1 0.91 0.89 0.79
To solve using trapezoidal rule

using trapezoidal rule.

To solve using Simpson’s rule

 using Simpson’s rule
Simpson’s rule is not applicable, as is not a multiple of 3

Example12 Evaluate up to 4 decimal places, using trapezoidal rule.

Solution: Taking , i.e.

0
2.71828 2.37744 1
By trapezoidal rule

using trapezoidal rule

Example13 From the following table, find the area bounded by the curve and
axis, between the ordinates to .
7.47 7.48 7.49 7.50 7.51 7.52
1.93 1.95 2.01 2.06
Solution: As , Simpson’s rule Simpson’s rules are not applicable.
Applying trapezoidal rule with

square units
Example14 The velocity of an airplane which starts from rest is given at fixed
intervals of time as shown:
(minutes) 2 4 6 8 10 12 14 16 18 20
8 17 24 28 30 20 12 6 2 0
(km/minutes)
Estimate the approximate distance covered in 20 minutes.
Solution: Since the airplane starts from rest, its initial velocity is zero. So the
time/velocity relationship may be tabulated as:
(minutes) 0 2 4 6 8 10 12 14 16 18 20
 0 8 17 24 28 30 20 12 6 2 0
(km/minutes)
Let be the distance covered at any instant of time ,
Then or
Distance covered in 20 minutes is given by:

Applying Simpson’s rule with , as

km
Example15 A solid of revolution is formed by rotating about axis, the area
between axis, the line and a curve through the points with the
following coordinates:
0 0.25 0.50 0.75 1
1 0.5846 0.5085
Estimate the volume of solid formed, giving the answer upto 3 decimal places.
Solution: Volume of solid formed by revolving the curve between the
ordinates and along is given by
By Simpson’s rule with , as