Approximation of the functions

(Part Two)

Jean Paul Nsabimana,

Department of Mathematics
University of Rwanda
Email: [email protected]

December 17, 2020

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
1 / 60
 1 Interpolation with Unequal Intervals
Newton’s Divided Difference Formula
Lagrange’s Interpolation Formula
2 Inverse Interpolation
Introduction
Lagrange’s Method
Iterative Method
3 Numerical Differentiation
Newton’s forward interpolation formula
Newton’s Backward interpolation formula
Maximum and Minimum values
4 Numerical Integration
Introduction
For n = 1: Trapezoidal Rule
For n = 2: Simpson 31 Rule
For n = 3: Simpson’s 38 Rule
Jean Paul Nsabimana, Department
Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
2 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

Let (xi , yi ), i = 0, 1, 2, 3, . . . , n be the set of values of x and y satisfying the

function y = f (x), the values of xi being not necessarily equally spaced.
The first divided difference of f (x) for the arguments x0 and x1 is defined
as f (xx11)−f
−x0
(x0 )
and denoted by [x0 , x1 ]. So that
f (x1 )−f (x0 )
f (x0 , x1 ) = [x0 , x1 ] = x1 −x0 .
f (x2 )−f (x1 )
Similarly, we have f (x1 , x2 ) = [x1 , x2 ] = x2 −x1 .
In general, we have

f (xn ) − f (xn−1 )
f (xn−1 , xn ) = [xn−1 , xn ] = .
xn − xn−1

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
3 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

Again, the second divided difference of f (x) for the arguments x0 , x1 and
x2 is defined as f (x0 , x1 , x2 ) = [x0 , x1 , x2 ] = [x1 , xx22]−[x
−x0
0 , x1 ]
.
[x2 , x3 ]−[x1 , x2 ]
f (x1 , x2 , x3 ) = [x1 , x2 , x3 ] = x3 −x1 .
The third divided difference is given as
f (x0 , x1 , x2 , x3 ) = [x0 , x1 , x2 , x3 ] = [x1 , x2 ,xx33]−[x
−x0
0 , x1 , x2 ]

In general, we have

f (x0 , x1 , . . . xn−1 ) = [x0 , x1 , . . . xn−1 ]

[x1 , . . . xn−1 ] − [x0 , x1 , . . . , xn−2 ]
= .
xn−1 − x0

It is the nth order divided difference of f (x) for x0 , x1 , . . . , xn−1 .

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
4 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

The divided difference table

x f (x) 1st 2nd 3rd
x0 f (x0 )
x1 f (x1 ) f (x0 , x1 )
x2 f (x2 ) f (x1 , x2 ) f (x0 , x1 , x2 )
x3 f (x3 ) f (x2 , x3 ) f (x1 , x2 , x3 ) f (x0 , x1 , x2 , x3 )
x4 f (x4 ) f (x3, x4 ) f (x2 , x3 , x4 ) f (x1 , x2 , x3 , x4 )

We note the following properties of divided differences:

f (x1 −f (x−0) f (x0 )−f (x1 )
[x0 , x1 ] = x1 −x0 = x0 −x1 = [x1 , x0 ].
[x1 , x2 ]−[x0 , x1 ] [x2 , x1 ]−[x1 , x0 ] [x1 , x0 ]−[x2 , x1 ]
[x0 , x1 , x2 ] = x2 −x0 = x2 −x0 = x0 −x2 ,
which shows that the divided difference is symmetric with respect to its
arguments.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
5 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

To derive a formula involving a function y = f (x) and its divided

differences, we observe that
y − y0
[x, x0 ] = [x0 , x] =
x − x0
⇒ y = y0 + (x − x0 )[x, x0 ] (∗)

[x0 , x1 ] − [x, x0 ] [x, x0 ] − [x0 , x1 ]

[x, x0 , x1 ] = =
x1 − x x − x1
⇒ [x, x0 ] = [x0 , x1 ] + (x − x0 )[x, x0 , x1 ] (∗∗)
Substituting (**) into (*), we obtain
y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 )[x, x0 , x1 ] (∗ ∗ ∗)
Further, we have
[x, x0 , x1 ] − [x0 , x1 , x2 ]
[x, x0 , x1 , x2 ] =
x − x2
⇒ [x, x0 , x1 ] = [x0 , x1 , x2 ] + (x − x2 )[x, x0 , x1 , x2 ] (∗ ∗ ∗∗)
Jean Paul Nsabimana, Department
Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
6 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

Putting (****) in (***), we obtain

y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 ) ([x0 , x1 , x2 ] + (x − x2 )[x, x0 , x1 ,

y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 )[x0 , x1 , x2 ]
+ (x − x0 )(x − x1 )(x − x2 )[x, x0 , x1 , x2 ].

Proceeding in this way, we obtain the formula

y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 )[x0 , x1 , x2 ]

+(x − x0 )(x − x1 )(x − x2 )[x, x0 , x1 , x2 ]
+ . . . + (x − x0 )(x − x1 ) . . . (x − xn )[x, x0 , x1 , . . . , xn ]. (i)

Where the last term is the reminder term in the formula. The formula (i)
above is called Newton’s Divided interpolation formula.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
7 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

Example
Construct a divided difference table for the following data.

x 4 5 7 10 11 13
f (x) 48 100 294 900 1210 2028

Solution: The divided difference table is given as follows

x f (x) 1st 2nd 3rd 4th

4 48
100−48
5 100 5−4 = 52
294−100 97−52
7 294 7−5 = 97 7−4 = 15
900−294 202−97 21−15
10 900 10−7 = 202 10−5 = 21 10−4 =1
1210−900 310−202 27−21
11 1210 11−10 = 310 11−7 = 27 11−5 =1 0
2028−1210 409−310 33−27
13 2028 13−11 = 409 13−10 = 33 13−7 =1 0

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
8 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Newton’s Divided Difference Formula

Using Newton’s divided difference formula, find the cubic function from the
following table of values:
x 0 2 3 5
y 1 15 40 156
Solution: We form the table of divided differences as follows:
x y 1st 2nd 3rd
0 1
2 15 7
3 40 25 6
5 156 58 11 1

y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 )[x0 , x1 , x2 ]

+ (x − x0 )(x − x1 )(x − x2 )[x0 , x1 , x2 , x3 ]
y = 1 + (x − 0)(7) + (x − 0)(x − 2)(6) + (x − 0)(x − 2)(x − 3)(1)
∴y = x 3 + x 2 + x + 1.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
Decemberof
17, Rwanda
2020 Em
9 / 60
 Interpolation with Unequal Intervals Newton’s Divided Difference Formula

Exercise II.3

1 Find a polynomial satisfied by the following data.

x -4 -1 0 2 5
f (x) 1245 33 5 9 1335

2 Find f (4) from the following table of values

x 0 1 3 5
f (x) 8 11 35 123

3 Find a third order polynomial satisfying the following date

x 0 2 3 6
f (x) 659 705 729 804

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
10 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Let y = f (x) be a function which takes the (n + 1) values y0 , y1 , y2 , . . . , yn

polynomial of nth degree in x. Let this polynomial be the form

y = f (x) = a0 (x − x1 )(x − x2 ) . . . (x − xn )
+ a1 (x − x0 )(x − x2 ) . . . (x − xn )
+ a2 (x − x0 )(x − x1 )(x − x3 ) . . . (x − xn ) + . . . (1)
+ an (x − x0 )(x − x1 ) . . . (x − xn−1 ).

Putting x = x0 , y = y0 in equation (2), we get

y0 = a0 (x0 − x1 )(x0 − x2 )(x0 − x3 ) . . . (x0 − xn )

y0
∴ a0 =
(x0 − x1 )(x0 − x2 )(x0 − x3 ) . . . (x0 − xn )

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
11 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Again putting x = x1 and y = y1 in equation (2), we get

y1
a1 = .
(x1 − x0 )(x1 − x2 )(x1 − x3 ) . . . (x1 − xn )

Proceeding in the same way, we obatin

y2
a2 = .
(x2 − x0 )(x2 − x1 )(x2 − x3 ) . . . (x2 − xn )

Until
yn
an = .
(xn − x0 )(xn − x1 )(xn − x2 ) . . . (xn − xn−1 )

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
12 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Substituting the values of a0 , a1 , a2 , . . . , an in equation (2), we get

(x − x1 )(x − x2 )(x − x3 ) . . . (x − xn )
y = f (x) = y0
(x0 − x1 )(x0 − x2 )(x0 − x3 ) . . . (x0 − xn )
(x − x0 )(x − x2 )(x − x3 ) . . . (x − xn )
+ y1
(x1 − x0 )(x1 − x2 )(x1 − x3 ) . . . (x1 − xn )
+ .......................................
(x − x0 )(x − x1 )(x − x2 ) . . . (x − xn−1 )
+ yn . (2)
(xn − x0 )(xn − x1 )(xn − x2 ) . . . (xn − xn−1 )

This is knwon as Lagrange’s Interpolation formula.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
13 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Note:
1 This formula can be used irrespectively of whether the values
x0 , x1 , x2 , . . . , xn are equally spaced or not.
2 It is simple and easy to remember but its application is not speedy.
3 The main drawback of it is that if another interpolation value is inserted,
then the interpolation coefficients are required to be recalculated.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
14 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Example
Use Lagrange’s interpolation formula to find the value of y when x = 10, if the
values of x and y are given as below:

x 5 6 9 11
y 12 13 14 16

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
15 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Lagrange’s Interpolation Formula

Solution: Here,
x0 = 5, x1 = 6, x2 = 9, x3 = 11, y0 = 12, y1 = 13, y2 = 14, y3 = 16
. By Lagrange’s formula in equation 2
(x − 6)(x − 9)(x − 11) (x − 5)(x − 9)(x − 11)
y= (12) + (13)
(5 − 6)(5 − 9)(5 − 11) (6 − 5)(6 − 9)(6 − 11)
(x − 5)(x − 6)(x − 11) (x − 5)(x − 6)(x − 9)
+ (14) + (16)
(9 − 5)(9 − 6)(9 − 11) (11 − 5)(11 − 6)(11 − 9)
Putting x = 10 in the above equation, we get
(10 − 6)(10 − 9)(10 − 11) (10 − 5)(x − 9)(10 − 11)
y10 = (12) + (13)
(5 − 6)(5 − 9)(5 − 11) (6 − 5)(6 − 9)(6 − 11)
(10 − 5)(10 − 6)(10 − 11) (10 − 5)(10 − 6)(10 − 9)
+ (14) + (16)
(9 − 5)(9 − 6)(9 − 11) (11 − 5)(11 − 6)(11 − 9)
∴ y10 = 14.6666667.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
16 / 60
 Interpolation with Unequal Intervals Lagrange’s Interpolation Formula

Exercise II.4

1 Use Lagrange’s interpolation formula to find the form of f (x), given

x 0 2 3 6
y 648 704 729 792

2 By means of Lagrange’s formula, prove that

y1 = y3 − 0.3(y5 − y−3 ) + 0.2(y−3 − y−5 )

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
17 / 60
 Inverse Interpolation Introduction

INVERSE INTERPOLATION

Definition
So far, given a set of values of x and y , we were required to find the value
of y corresponding to a value of x. Sometimes, we may require to find the
value of x corresponding to a certain value of y .
The process of finding such a value of x is called inverse interpolation.
In this section, we shall study two methods of inverse interpolation:
1 Lagrange’s method.
2 Iterative method.
We apply Lagrange’s method when the arguments (xi ) are unequally spaced
and the iterative method when the arguments are equally spaced.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
18 / 60
 Inverse Interpolation Lagrange’s Method

Lagrange’s Method

Choosing y as an independent variable and x and a dependent variable in

Lagrange’s interpolation.

(y − y1 )(y − y2 )(y − y3 ) . . . (y − yn )
x = f (y ) = x0
(y0 − y1 )(y0 − y2 )(y0 − y3 ) . . . (y0 − yn )
(y − y0 )(y − y2 )(y − y3 ) . . . (y − yn )
+ x1
(y1 − y0 )(y1 − y2 )(y1 − y3 ) . . . (y1 − yn )
+ .......................................
(y − y0 )(y − y1 )(y − y2 ) . . . (y − yn−1 )
+ xn . (3)
(yn − y0 )(yn − y1 )(yn − y2 ) . . . (yn − yn−1 )

Which is used for the inverse interpolation.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
19 / 60
 Inverse Interpolation Lagrange’s Method

Lagrange’s Method

Example
The following table gives the values of x and y :

x 30 35 40 45 50
y 15.9 14.9 14.1 13.3 12.5

Find the value of x corresponding to y = 13.6.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
20 / 60
 Inverse Interpolation Lagrange’s Method

Lagrange’s Method

Solution: Here, y0 = 15.9, y1 = 14.9, y2 = 14.1, y3 = 13.3, y4 = 12.5, x0 =

30, x1 = 35, x2 = 40, x3 = 45, x4 = 50. Taking y = 13.6 in Equation (3) above,
we get

(13.6 − 14.9)(13.6 − 14.1)(13.6 − 13.3)(13.6 − 12.5)

x = (30)
(15.9 − 14.9)(15.9 − 14.1)(15.9 − 13.3)(15.9 − 12.5)
(13.6 − 15.9)(13.6 − 14.1)(13.6 − 13.3)(13.6 − 12.5)
+ (35)
(14.9 − 15.9)(14.9 − 14.1)(14.9 − 13.3)(14.9 − 12.5)
(13.6 − 15.9)(13.6 − 14.9)(13.6 − 13.3)(13.6 − 12.5)
+ (40)
(14.1 − 15.9)(14.1 − 14.9)(14.1 − 13.3)(14.1 − 12.5)
(13.6 − 15.9)(13.6 − 14.9)(13.6 − 14.1)(13.6 − 12.5)
+ (45)
(13.3 − 15.9)(13.3 − 14.9)(13.3 − 14.1)(13.3 − 12.5)
(13.6 − 15.9)(13.6 − 14.9)(13.6 − 14.1)(13.6 − 13.3)
+ (50)
(12.5 − 15.9)(12.5 − 14.9)(12.5 − 14.1)(12.5 − 13.3)
∴x = 43.14185 =≈ 43.1.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
21 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Let us consider Netwoton’s forward interpolation formula:

x(x − 1) 2 x(x − 1)(x − 2) 3

yx = y0 + x4y0 + 4 y0 + 4 y0 + . . . (4)
2! 3!

From this we get

 
1 x(x − 1) 2 x(x − 1)(x − 2) 3
x= yx − y0 − 4 y0 − 4 y0 − . . . . (5)
4y0 2! 3!

Negleting the second and higher differences, we get a first approximation

for x and this we write as:
1
x (1) = [yx − y0 ] . (6)
4y0

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
22 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Substituting this in RHS of Equation (5), the second approximation is given

by
h i
x (1) (x (1) −1) 2 x (1) (x (1) −1)(x (1) −2) 3
x (2) = 4y1
0
y x − y 0 − 2! 4 y 0 − 3! 4 y 0 − . . . . (7)

The Third approximation is obtained by substituting x = x (2) in the RHS of

Equation (5).
h i
x (2) (x (2) −1) 2 x (2) (x (2) −1)(x (2) −2) 3
x (3) = 4y
1
0
y x − y 0 − 2! 4 y 0 − 3! 4 y 0 − . . . . (8)

Similarly, further approximations can be obtained ad the process us to be

continued until two successive approximations of x agree with other to the
required accuracy.
If the final approximation is x ∗ , then the required is x = x0 + x ∗ h.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
23 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Note:
1 This method can be applied equally well by starting with any other
interpolation formula.
2 It is powerful iterative procedure to find the root of an equation to a good
degree of accuracy.

Example
From the following table

x 1.8 2.0 2.2 2.4 2.6

Find x when y = 5, using iterative method.
y 2.9 3.6 4.4 5.5 6.7

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
24 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Solution:
The forward differences table is given as follows

x y 4 42 43 44
1.8 2.9
2.0 3.6 0.7
2.2 4.4 0.8 0.1
2.4 5.5 1.1 0.3 0.2
2.6 6.7 1.2 0.1 -0.2 -0.4

Here, yx = 5, y0 = 2.9, y1 = 3.6, y2 = 4.4, y3 = 5.5, y4 = 6.7, 4y0 =

0.7, 42 y0 = 0.1, 43 y0 = 0.2, 44 y0 = −0.4.
Let us consider Equation (6), we get

1 1
x (1) = [yx − y0 ] = (5 − 2.9) = 3.
4y0 0.7

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
25 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Solution:
Using Equation
h (7), then the second approximation is given as i
(1) (1) (1) (1) (1)
1
x (2) = 4y 0
y x − y0 − x (x2! −1) 42 y0 − x (x −1)(x
3!
−2) 3
4 y0 − . . .
 
1 3(3−1) 3(3−1)(3−2) 3(3−1)(3−2)(3−3)
= 0.7 5 − 2.9 − 2 (0.1) − 3! (0.2) − 4! (−0.4)
1
= 0.7 (5 − 2.9 − 0.3 − 0.2) = 2.2857143.

Proceeding on as explained above, we get

x (3) = 2.4724147
(4)
x = 2.64364724
x (5) = 2.5373869
(6)
x = 2.5985062
(7)
x = 2.5611579

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
26 / 60
 Inverse Interpolation Iterative Method

Iterative Method

Solution:
Proceeding on as explained above, we get
x (8) = 2.5841148
(9)
x = 2.5700665
(10)
x = 2.5786872
(11)
x = 2.573406
x (12) = 2.5766447
(13)
x = 2.5746598
(14)
x = 2.5758768
(15)
x = 2.575625.

Therefore ∴ x ∗ = 2.575, correct to three decimal places. Hence, the value

of x corresponding to to yx = 5 is
x = x0 + x ∗ h = 1.8 + (2.575)(0.2) = 2.315.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
27 / 60
 Inverse Interpolation Iterative Method

Exercise II.5

1 Given values f (14) = 68.7, f (17) = 64, f (31) = 44, and f (35) = 39. Find
f (27) using Lagrange’s formula.
2 Given u1 = 22, u2 = 30, u3 = 82, u4 = 106 and u5 = 206. Find u6 using
Lagrange’s interpolation formula.
3 Apply Lagrange’s formula inversely to obtain the root of f (x) = 0, given
that f (300 = −30, f (34) = −13, f (38) = 3, and f (42) = 18.
4 Find x when f (x) = 0.163, given that

x 80 82 84 86 88
f (x) 0.131 0.154 0.176 0.200 0.221

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
28 / 60
 Numerical Differentiation

Introduction

Numerical Differentiation Formula

We consider the general polynomial of interpolation which states that given
a set of values (xi , yi ), i = 0, 1, 2, 3, . . . of two variables x and y , we obtain
a polynomial φ(x), called interpolation polynomial, such that φ(x) agrees
with y (x) at all xi using Lagrange’s formula and Newton’s formula.
We shall be concerned with the problem of numerical differentiation which
states that given a set of values of x and y (x) as previously, we wish to
d 2y
derive formula for computing the approximate values of dy dx , dx 2 , . . . , for any
x ∈ [x0 , xn ].

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
29 / 60
 Numerical Differentiation

Newton’s forward interpolation formula

The general method for deriving the numerical differentiation formula is to

differentiate the interpolating polynomial. Hence, corresponding to each of
the formula derived in the section on interpolation; We may derive a
formula for the derivative.
We illustrate the derivation with Newton’s forward interpolation formula
only, the method of derivation being the same in each case.
Newton’s forward interpolation formula is
1 1
y = y0 + p4y0 + p(p − 1)42 y0 + p(p − 1)(p − 2)43 y0 + . . . ,(9)
2! 3!
x−x0
where x = x0 + ph, i.e p = h .

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
30 / 60
 Numerical Differentiation Newton’s forward interpolation formula

Newton’s forward interpolation formula

dy dy dp
Then dx = dp dx by Chain rule.

dy 1 1 1
= 4y0 + (2p − 1)42 y0 (10)
dx h 2! h
1 1
+ [(p − 1)(p − 2) + p(p − 1) + p(p − 2)] 43 + . . .
3! h

dy 1
h
2p−1 2 3p 2 −6p+2 3 4p 3 −18p 2 +22p−6 4
i
= h 4y0 + 2 4 y0 + 6 4 y0 + 24 4 y0 + . . . (11)
.
dx
(11)
Expression (11) can be used for computing the value of dy dx for any
non-tabular value of x in [x0 , xn ]. For tabular values of x, the formula takes
a simpler form.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
31 / 60
 Numerical Differentiation Newton’s forward interpolation formula

Newton’s forward interpolation formula

By setting x = x0 we obtain p = 0 from p = x−xh

0
and hence
 
dy
= h1 4y0 − 12 42 y0 + 13 43 y0 − 14 44 y0 + . . . .
 
(12)
dx x=x0

Differentiating (11) again, we obtain

d 2y h
12p 2 −36p+22 4
i
= 1
h 2 42 y0 + 6p−6 3
6 4 y0 + 24 4 y0 + ... . (13)
dx 2
From which we obtain
 2 
d y 1
 11 4

= h2 42 y0 − 43 y0 + 12 4 y0 + ... . (14)
dx 2 x=x0

Formula for computing higher derivatives may be obtained by successive

differentiation.
In a similar way, different differentiation formula can be derived by starting
with other interpolation formula.
Jean Paul Nsabimana, Department
Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
32 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Newton’s Backward Interpolation

Newton’s Backward Interpolation formula gives

   
dy 1 1 2 1 3
= ∇yn + ∇ yn + ∇ yn + . . . . (15)
dx x=xn h 2 3

And
d 2y
   
1 2 3 11 4 5 5
= ∇ y n + ∇ y n + ∇ y n + ∇ y n + . . . . (16)
dx 2 x=xn h2 12 6

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
33 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Example

Example
The population of a certain town (as obtain from census data) is shown in the
following table

year (x) 1951 1961 1971 1981 1991

Population in thousands (y) 19.96 36.65 58.81 77.21 94.61

Find the rate of growth of the population in the year 1981.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
34 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Solution

We have to find the derivative at 1981 which is near the end of the table. Hence,
we use the derivative of Newton’s backward interpolation formula. The table of
differences is given as follows

x y ∇y ∇2 y ∇3 y ∇4 y
1951 19.96
1961 36.65 16.69
1971 58.81 22.16 5.47
1981 77.21 18.40 -3.76 -9.23
1991 94.61 17.40 -1 2.76 11.99

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
35 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Solution

Here, h = 10, xn = 1991, ∇yn = 17.4, ∇2 yn = −1, ∇3 yn = 2.76, ∇4 yn = 11.99.

So,
dy h 2 3 2
i
= h1 ∇yn + 2p+1 2 ∇2 yn + 3p +6p+2
6 ∇3 yn + 2p +9p12+11p+3 ∇4 yn .
dx
Now, we have to find out the rate of growth of the population in the year 1981.
Then,
1981 − xn 1981 − 1991
xn + ph = 1981 ⇒ p = = = −1.
h 10
h i
dy
dx =
 x=1981 
1 2(−1)+1 3(−1)2 +6(−1)+2 2(−1)3 +9(−1)2 +11(−1)+3
10 17.4 + 2 (−1) + 6 (2.76) + 24 (11.99)
 
dy
= 1.6440833.
dx x=1981

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
36 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Example

Example
dy d 2y
Find the values of dx and dx 2 at x = 1.2 from the following table.

x 1.0 1.2 1.4 1.6 1.8 2.0 2.2

y 2.7183 3.3201 4.0552 4.9530 6.0496 7.3891 9.025

Solution: The difference table is

x y 4y 42 y 43 y 44 y 45 y 46 y
1.0 2.7183
1.2 3.3201 0.6018
1.4 4.0552 0.7351 0.1333
1.6 4.9530 0.8978 0.1627 0.0294
1.8 8.0496 1.0966 0.1988 0.0361 0.0067
2.0 7.3891 1.3395 0.2429 0.0441 0.0080 0.0013
2.2 9.025 0.2964 0.2964 0.0535 0.0094 0.0014 0.0001

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
37 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Solution

Here,
 x = 1.2, h = 0.2, p = x−x
h
0
= 1.2−1
0.2 = 1.
 2

dy  3p −6p+2 3 2p 3 −9p 2 +11p−3 4
= h1 4y0 + 2p−1
 2
dx  2 4 y0 + 6 4 y0 + 12 4 y0 + ...
 x=1.2
dy 

dx  =
x=1.2
1
0.618 + 21 (0.1333) − 61 (0.0294) + 1 1 1


0.2 12 (0.0067) + 20 (0.0013) + 30 (0.0001)
dy 

dx  = 3.2305.
x=1.2
On the other hand, by setting x0 = 1.2 and y0 = 3.3201, we get
  
dy  1 1 2 1 3 1 4
= 4y 0 − 4 y 0 + 4 y 0 − 4 y 0 + . . .
dx x=x0 h 2 3 4

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
38 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Solution Cont....

  
dy 
 1
dx  = 0.2 3.3201 − 12 (0.1627) + 13 (0.0361) − 1
4(0.0080)+ 15 (0.0014)
x=1.2

dy 
= 3.3205
dx x=1.2

d 2 y 
  
1 2 11 4
3 5 5
= 4 y0 − 4 y0 + 4 y0 − 4 y0 + . . .
dx 2 x=x0 h2 12 6
d 2 y 
  
1 11 5
= 0.1627 − 0.361 + (0.0080) − (0.0014)
dx 2  x=1.2 0.22 22 6
= 3.318.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
39 / 60
 Numerical Differentiation Newton’s Backward interpolation formula

Exercise II.6

1 The table given below reveals velocity v of a body during time t specified.
Find its acceleration at t = 1.1.
t 1.0 1.1 1.2 1.3 1.4
v 43.1 47.7 52.1 56.4 60.8

2 For the following pairs of values of x and y , find numerically the first
derivative at x = 4.
x 1 2 4 8 10
y 0 1 5 21 27

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
40 / 60
 Numerical Differentiation Maximum and Minimum values

Maximum and Minimum values of a tabulated function

The maximum and minimum values of a function can be found by equating

the first derivative to zero and solving the variable.
The same procedure can be applied to determine the maximum and
minimum of a tabulated function.
We consider Newton’s forward difference formula:
p(p − 1) 2 p(p − 1)(p − 2) 3
y = y0 + p4y0 + 4 y0 + 4 y0 + . . .
2 6

Differentiating with respect to p, we obtain

dy 2p − 1 2 3p 2 − 6p + 2 3
= 4y0 + 4 y0 + 4 y0 + . . .
dp 2 6

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
41 / 60
 Numerical Differentiation Maximum and Minimum values

Maximum and Minimum values of a tabulated function

dy
For maxima or minima, dp = 0.
Determination the right-hand side after third order difference and equating
it to zero, we obtain
 
1 2 1 3
1
4y0 − 4 y0 + 4 y0 + 42 y0 − 43 y0 + (42 y0 )p 2 = 0
2 3 2

which is a quadratic equation in p and can be solved. The values of x can

be found from the relation
x = x0 + ph.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
42 / 60
 Numerical Differentiation Maximum and Minimum values

Example

Example
Given the following data, find the maximum value of y .

x -1 1 2 3
y -21 15 12 3

Solution: Since the arguments are not equispaced, we will form the divided
difference table as follows
x y 1st divided difference 2nd div. differ 3rd div. differ
-1 -21
1 15 18
2 12 -3 -7
3 3 -9 -3 1

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
43 / 60
 Numerical Differentiation Maximum and Minimum values

Solution Cont...

Using Newton’s divided difference formula, we get

y = y0 + (x − x0 )[x0 , x1 ] + (x − x0 )(x − x1 )[x0 , x1 , x2 ] + (x − x0 )(x −
x1 )(x − x2 )[x0 , x1 , x2 , x3 ]
y = −21 + (x + 1)(18) + (x + 1)(x − 1)(−7) + (x + 1)(x − 1)(x − 2)(1)
y = x 3 − 9x 2 + 17x + 6
2
For maximum, dydx = 0 and
d y
dx 2 < 0. That is
( (
3x 2 − 18x + 17 = 0 x = 4.8257 or x = 1.743
⇒ ⇒ x = 1.743
6x − 18 < 0 x <3

Thus, the maximum value is

y (1.743) = (1.743)3 − 9(1.743)2 + 17(1.743) + 6 = 15.171612.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
44 / 60
 Numerical Differentiation Maximum and Minimum values

Example

Determine the maxima and minima of the function y = f (x) tabulated below.

x 0 1 2 3
y -5 -7 -3 13

Solution: The difference table is

x y 4 42 43
0 -5
1 -7 -2
2 -3 4 6
3 13 16 12 6
x−x0
Here, x0 = 0, h = 1, then p = h = x.
p(p−1) 2 p(p−1)(p−2) 3
Now, y = y0 + p4y0 + 2 4 y0 + 6 4 y0 + ...,

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
45 / 60
 Numerical Differentiation Maximum and Minimum values

Solution Cont....

Hence
x(x − 1) x(x − 1)(x − 2)
y = −5 + x(−2) + (6) + (6)
2 6
y = x 3 − 3x − 5.

In order to find the maxima and minima of the function, we have to solve
the following equations
( (
dy dy 2
dx = 0 dx = 3x − 3 = 0
2
d y ⇒ 2
d y ⇒ x = ±1
dx 2 dx 2 = 6x

So there is a maxima at x = −1 and a minima at x = 1.

y (−1) = (−1)3 − 3(−1) − 5 = −3 Maxima

3
y (1) = (1) − 3(1) − 5 = −7 Minima

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
46 / 60
 Numerical Differentiation Maximum and Minimum values

Exercise II.7

1 Find the maximum and minimum values of the function from the following
table
x 0 1 2 3 4 5
y 0 0.25 0 2.25 16.00 56.25
dy d 2y
2 Evaluate dx and dx 2 for the following table

x 0 1 3 6
y 18 10 -18 90
3 For the following data gives corresponding values of pressure and specific
volume of super-heated steam.
v 2 4 6 8 10
p 105 42.07 25.3 16.7 13
Find the rate of change of pressure w.r.t volume when v = 2 and the rate
of change of volume w.r.t pressure when p = 105.
Jean Paul Nsabimana, Department
Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
47 / 60
 Numerical Integration Introduction

Numerical Integration

We are required to compute the value of the definite integral

Z b
I = y dx, (17)
a

where the function y = f (x) is defined by the set of data points

(xi , yi ), i = 0, 1, 2, 3, . . . , n.
To derive a general formula for numerical integration, we substitute
Newton’s forward difference formula for y (x) in Equation (17), where the
interval [a, b] is divided into n equal sub-intervals such that
a = x0 < x1 < x2 < . . . < xn = b.
We obtain
Rx 
I = x0n y0 + p4y0 + 12 p(p − 1)42 y0 + 16 p(p − 1)(p − 2)43 y0 + . . . dx,

(18)

where putting x = x0 + ph. So dx = h dp, we get

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
48 / 60
 Numerical Integration Introduction

Numerical Integration

where putting x = x0 + ph. So dx = h dp, we get

Z n  
1 1
I = h y0 + p4y0 + p(p − 1)42 y0 + p(p − 1)(p − 2)43 y0 + . . . dp
0 2 6
2 2
n2 (n − 2)2 3
 
n n (2n − 3) 2
= h ny0 + 4y0 + 4 y0 + 4 y0 + . . . .
2 12 24

Therefore,
Z xn h i
n(2n−3) 2 n(n−2)2
y dx = nh y0 + n2 4y0 + 12 4 y0 + 24 43 y0 + . . . . (19)
x0

The equation (19) is the general quadratic formula and is called

Newton’s-cote’s formula.
Different numerical integration can be derived by setting n = 1, 2, 3, . . . , in
Equation (19)

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
49 / 60
 Numerical Integration For n = 1: Trapezoidal Rule

Trapezoidal Rule

We define that differences of second and higher order vanish, and we obtain
Z x1  
1 h h
y dx = h y0 + 4y0 = (2y0 + 4y0 ) = (y0 + y1 )
x0 2 2 2

For subsequent interval, we similarly obtain

Z x2 Z x3
h h
y dx = (y1 + y2 ); y dx = (y2 + y3 )
x1 2 x2 2
R xn
Until xn−1
y dx = h2 (yn−1 + yn ).
Combining the above formula, we obtain the rule
Z xn
h
y dx = [y0 + 2(y1 + y2 + y3 + . . . + yn−1 ) + yn ] , (20)
x0 2

which is known as the Trapezoidal rule.

Jean Paul Nsabimana, Department
Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
50 / 60
 Numerical Integration For n = 2: Simpson 31 Rule

1
Simpson 3 Rule

Setting n = 2 in Equation (19), we observe that differences of third and

higher order vanish, and we obtain
Z x2  
2 2(2.2 − 3 2
y dx = 2h y0 + 4y0 + 4 y0
x0 2 12
 
1 2
= 2h y0 + 4y0 + 4 y0
6
h
6y0 + 64y0 + 42 y0

=
3
h
= [6y0 + 6(y1 − y0 ) + (y2 − 2y1 + y0 )]
Z x2 3
h
∴ y dx = [y0 + 4y1 + y2 ] .
x0 3

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
51 / 60
 Numerical Integration For n = 2: Simpson 31 Rule

1
Simpson 3 Rule

Similarly,
Z x4
h
y dx = [y2 + 4y3 + y4 ] ; . . .
x2 3
Z xn
h
y dx = [yn−2 + 4yn−1 + yn ] .
xn−2 3

Summing up the above results, we obtain

Z xn
h
y dx = [y0 + 4(y1 + y3 + . . . + yn−1 )]
x0 3
h
+ [2(y2 + y4 + . . . + yn−2 ) + yn ] , (21)
3
1
which is known as Simpson’s 3 rule, and requires an even number of
sub-intervals of width h.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
52 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

3
Simpson’s 8 Rule

Setting n = 3 in Equation (19) and negleting all differences above the third
order, we get
Z x3  
3 3 2 1 3
y dx = 3h y0 + 4y0 + 4 y0 + 4 y0
x0 2 4 8
3h
= [y0 + 3y1 + 3y2 + y3 ] .
8
Similarly,
Z x6
3h
y dx = [y3 + 3y4 + 3y5 + y6 ]
x3 8
.
.
.
Z xn
3h
y dx = [yn−3 + 3yn−2 + 3yn−1 + yn ]
xn−3 8

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
53 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

3
Simpson’s 8 Rule

Adding up all these integrals, we obtain

Z xn
3h
y dx = [y0 + 3(y1 + y2 + y4 + y5 + . . . + yn−2 + yn−1 )]
x0 8
3h
+ [2(y3 + y6 + . . . + yn−3 ) + yn ] ,
8
(22)

which is called the Simpson’s 38 rule.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
54 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Example

Evaluate Z 10
dx
by using
0 1 + x2

1 Trapezoidal rule.
1
2 Simpson’s 3 rule.
3
3 Simpson’s 8 rule.
Solution: Dividing the whole range of integration [0, 10] into 10 equal parts;
h = 1 and the value of integral for each point of subdivision are given below.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
55 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Solution

x y
x0 = 0 y0 = 1
x1 = 1 y1 = 12 = 0.5
x2 = 2 y2 = 15 = 0.2
1
x3 = 3 y3 = 10 = 0.1
1
x4 = 4 y4 = 17 = 0.0588235
1
x5 = 5 Y5 = 26 = 0.0384615
1
x6 = 6 y6 = 37 = 0.027027
1
x7 = 7 y7 = 50 = 0.02
1
x8 = 8 y8 = 65 = 0.0153846
1
x9 = 9 y9 = 82 = 0.0121951
1
x10 = 10 y10 = 101

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
56 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Solution

1 By Trapezoidal Rule
Z 10
h
y dx = [y0 + 2(y1 + y2 + y3 + . . . + y9 ) + y10 ] = 1.4768422.
0 2

2 By Simpson’s 13 rule
R 10
0
y dx = h3 [y0 + 4(y1 + y3 + . . . + y9 ) + 2(y2 + y4 + . . . + y8 ) + y10 ]
Z 10
∴ y dx = 1.4316659.
0

3 Simpson’s 83 rule
R 10
0
y dx = 3h
8 [y0 + 3(y1 + y2 + y4 + y5 + y7 + y8 ) + 2(y3 + y6 + y9 ) + y10 ],
Z 10
∴ y dx = 1.4198828.
0

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
57 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Example

The velocity v of a particle at distance s from a point on its path is given by the
following table:

s[ft] 0 10 20 30 40 50 60
v [t/s] 47 58 64 65 61 52 38

Estimate the time taken to travel 60ft using Simpson’s 13 rule. Compare the
result with Simpson’s 38 rule.
Solution:
ds
v= = rate of displacement is velocity
dt
R 60 R 60
So , t = 0 v1 ds = 0 y dx = time taken to travel 60 ft. It means that y = 1
v
and s = x.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
58 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Solution

s[x] y = 1/v
1
x0 = 0 y0 = 47 = 0.0212765
1
x1 = 10 y1 = 58 = 0.0172413
1
x2 = 20 y2 = 64 = 0.015625
1
x3 = 30 y3 = 65 = 0.0153846
1
x4 = 40 y4 = 61 = 0.0163934
1
x5 = 50 y5 = 52 = 0.192307
1
x6 = 60 y6 = 38 = 0.0263157
Here h = 10; By simpson’s 31 rule, we get
Z 60
h
y dx = (y0 + 2(y2 + y4 ) + 4(y1 + y3 + y5 ) + y6 ) = 1.063502.
0 3
By Simpson’s 38 rule, we get
Z 60
3h
y dx = (y0 + 3(y1 + y2 + y4 + y5 ) + 2(y3 ) + y6 ) = 1.0643723.
0 8

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
59 / 60
 Numerical Integration For n = 3: Simpson’s 83 Rule

Exercise II.8

R2
1 Evaluate 0
y dx from the following table using Trapezoidal rule

x 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

y 1.21 1.37 1.46 1.59 1.67 2.31 2.91 3.83 4.01 4.79

2 When an a train is moving at 30 miles an hour, steam is shut off and brakes
are applied. The speed of the train in miles per hour after t seconds is
given by

t 0 5 10 15 20 25 30 35 40
v 30 24 19.5 16 13.6 11.7 10.0 8.5 7.0

Determine how far the train has moved in the 40 seconds.

Jean Paul Nsabimana, Department

Approximation
of Mathematics
of the functions
University
(Part Two)
December of 2020
17, Rwanda Em
60 / 60