MA206L-Numerical Methods

Lecture 18: Numerical Integration - Newton-Cotes

Panchatcharam Mariappan1

1 Associate Professor

Department of Mathematics and Statistics

IIT Tirupati, Tirupati

September 24, 2024

Open Newton-Cotes
Formula

1
Recap
 b n
X
f (x)dx ≈ Ai f (xi ) (1)
a i=0
 
b−a ad − bc
λ : [c, d] → [a, b], λ(x) = x+
d−c d−c
 b  d
b−a
f (x)dx = f (λ(u))du
a d−c c
 b  1
λ : [0, 1] → [a, b], λ(x) = (b − a)x + a =⇒ f (x)dx = (b − a) f (λ(u))du (2)
a 0

 b  1
1 1 1
λ : [−1, 1] → [a, b], λ(x) = (b − a)x + (a + b), f (x)dx = (b − a) f (λ(u))du
2 2 a 2 −1
(3)
2
Recap
Trapezoidal Rule:
 b
h
f (x)dx ≈ [f (a) + f (b)]
a 2
Simpson’s 1/3rd Rule
 b
h
f (x)dx ≈ [f (a) + 4f (a + h) + f (b)]
a 3

Simpson’s 3/8rd Rule

 b
3h
f (x)dx ≈ [f (a) + 3f (a + h) + 3f (a + 2h) + f (b)]
a 8

3
Open Newton-Cotes Formula
• The open Newton-Cotes formula do not include the endpoints [a, b] as
nodes.
• They use only interior nodes.
• The numerical approximation for open Newton-Cotes formula is same as
(1), but, xi ∈ (a, b).
• In order to use the same formula (1), we redefine the points by spacing as
follows: x−1 = a, x0 = a + h, xn+1 = b, xi = x0 + ih for each
i = 0, 1, 2, · · · , n where h = (b − a)/(n + 2).

4
Midpoint Rule
For the midpoint formula, we take n = 0, then x−1 = a, x0 = (a + b)/2, x1 = b.
Therefore, the interior point is only x0 = (a + b)/2. When we work on the
interval [0, 1], change of intervals and method of undetermined coefficients,
we obtain that 1  
1
f (x)dx ∼ A
= 0 f (4)
0 2
By the method of undetermined coefficients, we evaluate them on 1.
 1
f (x) = 1 : dx = A0 =⇒ A0 = 1
0

5
Midpoint Rule
The resulting formula is  1  
1
f (x)dx ≈ f (5)
0 2
By (2), we obtain that
 b  1   
1
f (x)dx = (b − a) f (λ(u))du ∼
= (b − a)A0 f λ
a 0 2
 
a+b
≈ (b − a)f
2

If h = (b − a)/2, then
 b
f (x)dx ≈ 2hf (a + h)
a
6
Two-Point Newton-Cotes Open Rule
For the Two-point formula, we take n = 1, then
x−1 = a, x0 = (a + b)/3, x0 = 2(a + b)/3, x2 = b. Therefore, the interior points
are only x0 = (a + b)/3 and x1 = 2(a + b)/3. When we work on the interval
[0, 1], change of intervals and method of undetermined coefficients, we obtain
that 1    
∼ 1 2
f (x)dx = A0 f + A1 f (6)
0 3 3

7
Two-Point Newton-Cotes Open Rule
By the method of undetermined coefficients, we evaluate them on {1, x}.
 1
f (x) = 1 : dx = A0 + A1 =⇒ A0 + A1 = 1

0
1
1 2 1 2 1
f (x) = x : xdx = A0 + A1 =⇒ A0 + A1 =
0 3 3 3 3 2

Solving for A0 , A1 , we obtain that A0 = 1/2, A1 = 1/2.

8
Two-Point Newton-Cotes Open Rule
The resulting formula is
 1    
1 1 1 2
f (x)dx ≈ f + f (7)
0 2 3 2 3

By (2), we obtain that

 b  1
f (x)dx = (b − a) f (λ(u))du
a 0
     
∼ 1 2
= (b − a)A0 f λ + (b − a)A1 f λ
3 3
    
(b − a) b−a b−a
≈ f a+ +f a+2
2 3 3

9
Two-Point Newton-Cotes Open Rule
If h = (b − a)/3, then
 b
3h
f (x)dx ≈ [f (a + h) + f (a + 2h)]
a 2

10
Error in Midpoint and Two-Point Newton-Cotes
Open Rule

Theorem 1 (Error in Midpoint Rule)

If f ∈ C 2 [a, b], then the error in Midpoint rule is given by

(b − a)3 ′′
f (ξ)
24
for some ξ ∈ (a, b)
Theorem 2 (Error in Two-point Newton-Cotes Open Rule)
If f ∈ C 2 [a, b], then the error in Two-point rule is given by

(b − a)3 ′′
f (ξ)
36
for some ξ ∈ (a, b)
11
Midpoint Rule

Example 3
Evaluate  1
2
e−x dx
0
using Midpoint rule.
 
a+b a+b
a = 0, b = 1, = 0.5, f = 0.7788
2 2
 1
2
e−x dx = (1 − 0)f (0.5) = 0.7788
0

12
Two-Point Newton-Cotes Open Rule
Example 4
Using Two-Point Newton-Cotes open rule, the following numerical integration
can be obtained.
 2
2
x2 dx = 3−0 4
2 [9 + 16
9 ] = 10
3 , Et =
3
 3
0
1 3−0
dx = 2 [0.5 + 13 ] = 1.25, Et = 0.1363
1+x
 3p 0
√ √
3−0
1 + x2 dx = 2 [ 2+ 5] = 5.0673, Et = −0.5853
0
 3
ex dx = 3−0
2 [e + e2 ] = 15.16, Et = 3.9246
0

13
Some Open
Newton-Cotes Formula

14
Some Open Newton-Cotes Formula
When n = 0, open Newton-Cotes formula provides Midpoint rule. Similarly,
when n = 1, it yields the two-point Newton-Cotes open rule. When n = 3, the
following three point Newton-Cotes open rule is obtained
 b
4h
f (x)dx ≈ [2f (a + h) − f (a + 2h) + 2f (a + 3h)]
a 3

with error as
14 5 (4)
E3 (t) = h f (ξ), ξ ∈ (a, b)
45

15
Some Open Newton-Cotes Formula
When n = 4, we obtain the following four-point Newton-Cotes open rule
 b
5h
f (x)dx ≈ [11f (a + h) + f (a + 2h) + f (a + 3h) + 11f (a + 4h)]
a 24

where the error is

95 5 (4)
E4 (t) = h f (ξ), ξ ∈ (a, b)
144

16
Some Open Newton-Cotes Formula
When n = 5, we obtain the following five-point Newton-Cotes open rule
 b
6h
f (x)dx ≈ [11f (a + h) − 14f (a + 2h) + 26f (a + 3h) − 14f (a + 4h) + 11f (a + 5h)]
a 20

where the error is

41 7 (6)
h f (ξ), ξ ∈ (a, b)
E5 (t) = −
140
The following theorem gives the general error term for an (n + 1)-points
closed Newton-Cotes formula.

17
Some Open Newton-Cotes Formula

Theorem 5
Suppose that (1) denotes the open Newton-Cotes formula with x−1 = a, xn+1 =
b, x0 = a + h, xi = x0 + ih, i = 0, 1, 2, · · · , n where h = (b − a)/(n + 2). There
exists ξ ∈ (a, b) for which
 b n
X
f (x)dx = Ai f (xi ) + En (t)
a i=0

where the error is given by

 n+3  n+1
h (n+2)
f (ξ) t2 (t − 1) · · · (t − n)dt n is even, f ∈ C n+2


(n + 2)!

En (t) =  n+1
−1
hn+2 (n+1)
f (ξ) t(t − 1) · · · (t − n)dt n is odd, f ∈ C n+1



(n + 1)! −1
18
Composite Rules

19
Composite Rules
• The Newton-Cotes formulas are generally unsuitable for use over large
integration intervals.
• High-degree formulas would be required and the values of these
formulas are difficult to obtain.
• Also, the Newton-Cotes formulas are based on interpolation polynomials
that use equally spaced nodes, a procedure that is inaccurate over large
intervals because of the oscillatory nature of high-degree polynomials,
• Remedy: a piecewise approach to numerical integration that uses the
low-order Newton-Cotes formulas.
• These are the techniques most often applied.

20
Composite Rules

Theorem 6
Suppose that c ∈ [a, b], then
 b  c  b
f (x)dx = f (x)dx + f (x)dx
a a c

If we subdivide the interval into n equal subintervals by points

x0 = a, x1 = a + h, x2 = a + 2h, · · · , xi = a + ih, · · · , xn = b with spacing
h = (b − a)/n. Then by integration property, we have
 b X  xi+1
n−1
f (x)dx = f (x)dx
a i=0 xi

21
Composite Rules
Then we can apply the basic rules in each subinterval [xi , xi+1 ], that is,
 xi+1 m
X
f (x)dx = Aj f (xj )
xi j=0

where m is the number of points considered in the Newton-Cotes rule in the

interval [xi , xi+1 ]. The general composite Newton-Cotes formula is given by
 b n−1
XX m
f (x)dx = Aij f (xij ) (8)
a i=0 j=0

22
Composite trapezoidal Rule
The composite trapezoidal rule is obtained by keeping the m = 2 in (8). That
is,  xi+1
h
f (x)dx ≈ [f (xi ) + f (xi+1 )]
xi 2
and
 b n−1
hX
f (x)dx ≈ [f (xi ) + f (xi+1 )] (9)
a 2
i=0

23
Composite trapezoidal Rule
For each interval [xi , xi+1 ], the error is given by

h3 ′′
Eti = − f (ξi )
12
Therefore, the error of the composite trapezoidal rule is obtained as
n−1
h3 X ′′
Et = − f (ξi ) = O(h2 )
12
i=0

The last equation is obtained by using the reason that

n−1 n−1
h3 X ′′ h2 X1 b − a 2 ′′
− f (ξi ) = − (b − a) f ′′ (ξi ) = − h f (ζ)
12 12 n 12
i=0 i=0

for some ζ ∈ (a, b). 24

Composite trapezoidal Rule
The composite trapezoidal rule can also be written as
 b n−1
X h
f (x)dx ≈ h f (xi ) + [f (a) + f (b)] (10)
a 2
i=1

25
Composite trapezoidal Rule
Example 7
If the composite trapezoidal rule is to be used to compute
 1
2
e−x dx
0

with an error of at most 1

2 × 10−4 , how many points should be used?
Solution The error formula is
b − a 2 ′′
− h f (ζ)
12
Now,
2
f ′′ (x) = (4x2 − 2)e−x
Therefore,
|f ′′ (ζ)| ≤ 2 26
Composite trapezoidal Rule
To have an error of at most 1
2 × 10−4 , we must choose h such that

1−0 2 1
− h 2 ≤ × 10−4 =⇒ h ≤ 0.01733
12 2
As h = 1/n, we need at least 59 or more points to obtain the desired accuracy.

27
Composite Simpson’s 1/3 Rule
The composite Simpson’s 1/3 rule is obtained by keeping the m = 3 in (8).
 xi+1  
h xi + xi+1
f (x)dx ≈ [f (xi ) + 4f + f (xi+1 )]
xi 3 2

28
Composite Simpson’s 1/3 Rule
*Now choose n such that n is divisible by 2, then
 b n−1  
hX xi + xi+1
f (x)dx ≈ [f (xi ) + 4f + f (xi+1 )] (11)
a 3 2
i=0

For each interval [xi , xi+1 ], the error is given by

h5 (4)
Eti = − f (ξi )
90

29
Composite Simpson’s 1/3 Rule
Therefore, the error of the composite Simpson’s rule is obtained as
n−1
h5 X (4)
Et = − f (ξi ) = O(h4 )
90
i=0

When n is divisible by 2, the composite Simpson’s 1/3 rule can also be written
as
 b n/2 (n−2)/2
X X h
f (x)dx ≈ 4 [f (a+ (2i−1)h)] + 4 [f (a+ 2ih)] + [f (a) +f (b)] (12)
a 3
i=1 i=1

30
Composite Simpson’s 1/3 Rule

Example 8
The calculation of work is an important equation in science and engineering.
The general formula is given by

Work = Force × Distance

Although, this formula seems to be simple, when apply this to realistic problem,
we obtain a complicated function. For example, when the force varies during
the course of calculation which results in
 xn
W = F (x)dx
x0

31
Composite Simpson’s 1/3 Rule
When the angle between the force and direction of the movement also varies
as a function of position, we obtain that
 xn
W = F (x) cos(θ(x))dx
x0

The following table shows the force and angle as a function of position x.
x 0 5 10 15 20 25 30
F (x), N 0.0 9.0 13.0 14.0 10.5 12.0 5.0
θ(x), rad 0.50 1.40 0.75 0.90 1.30 1.48 1.50

Evaluate W using the Simpson’s 1/3 rule and composite Simpson’s 1/3 rules

32
Composite Simpson’s 1/3 Rule
Solution:
By Simpson’s rule, we have h = (30 − 0)/2
15
W = [F (0) cos(θ(0)) + 4F (15) cos(θ(0.9)) + F (30) cos(θ(1.50))]
3
= 5[0 + 34.81 + 0.3537] = 175.82
By composite Simpson’s 1/3 rule, we have h = (30 − 0)/6
5
W = [F (0) cos(θ(0)) + 4F (5) cos(θ(1.4)) + F (10) cos(θ(0.75))]
3
5
+ [F (10) cos(θ(0.75)) + 4F (15) cos(θ(0.90)) + F (20) cos(θ(1.30))]
3
5
+ [F (20) cos(θ(1.30)) + 4F (25) cos(θ(1.48)) + F (30) cos(θ(1.50))]
3
= 117.13
33
Composite Midpoint Rule
Let f ∈ c2 [a, b], n be even, h = (b − a)/(n + 2) and xi = a + (i + 1)h for each
i = −1, 0, · · · , n + 1. Then the composite midpoint rule is given by
 b n/2
X
f (x)dx ≈ 2h f (x2i ) (13)
a i=0

The error of the composite midpoint rule is given by

b − a 2 ′′
h f (ζ), ζ ∈ (a, b)
6

34
Merits and Demerits of
Newton-Cotes Rules

35
Merits and Demerits
• Simpson’s 1/3 rule has less truncation error compare to Simpson’s 3/8
rule.
• However, 3/8 rule requires 4 points and 1/3 rule requires only 3 points for
the evaluation.
• In composite rule, with (6n + 1) points, the truncation error of 1/3 rule will
be better than 3/8 rule.
• In general, Newton-Cotes formula has no advantage on even number of
points.

36
Merits and Demerits
• Besides the formulas with even number of points are useful for
integrating tabulated functions.
• Simpson’s 1/3 rule is usually the method of preference because it attains
third order accuracy with three points rather than the four points required
for the 3/8 version.
• However, the 3/8 rule has utility when the number of segments is odd.
• We can use Simpson’s rule to integrate the function for four segments.

37
Merits and Demerits
• Suppose that you desired an estimate for five segments. One option
would be composite trapezoidal rule.
• This may not be advisable, however, because of the large truncation error
associated with this method.
• An alternative would be to apply Simpson’s 1/3 rule to the first two
segments and Simpson’s 3/8 rule to the last three
• In this way, we could obtain an estimate with third-order accuracy across
the entire interval.
• One of the main disadvantage of composite Simpson’s 1/3 rule is that an
even number of intervals must be utilized to implement the method.

38
Merits and Demerits
• If the function is known only in table form, it is most convenient to use
Newton-Cotes formula.
• The composite trapezoidal rule is most convenient for even when the
table contains unequally spaced points.
• Notice that, as was the case with Simpson’s 1/3 and 3/8 rules, the five
and six point rules have the same order error.
• This general characteristic holds for the higher point rules and leads to
the result that the even-segment–odd-point formulas (for example, 1/3
rule and Boole’s rule) are usually the methods of preference.
• However, it must also be stressed that, in engineering practice, the
higher-order (that is, greater than four-point) formulas are rarely used.

39
Merits and Demerits
• Simpson’s rules are sufficient for most applications.
• Accuracy can be improved by using the composite rules.
• Newton-Cotes rule uses evenly spaced points in the interval.
• It uses a polynomial of minimal degree and exploits the Lagrange
interpolation.
• When Newton-Cotes formula uses n nodes, it is exact for polynomials of
degree at most n − 1.
• Since higher order methods are based on Lagrange interpolation, they
suffer from ill-conditioned.
• It can produce negative Ai .

40
Merits and Demerits
• It has limited applications due to failure of some function where the
problem arises from the assumption that the interpolation points are
uniformly spaced.
• Alternatively, the Gaussian quadrature does not require uniformly spaced
points as discussed in the next chapter.
• Furthermore, when the function is known and high accuracy is required,
methods such as Romberg integration or Gauss quadrature are
recommended.

41
 Thanks
Doubts and Suggestions
[email protected]

42
MA206L-Numerical Methods
Lecture 18: Numerical Integration - Newton-Cotes

Panchatcharam Mariappan1

1 Associate Professor

Department of Mathematics and Statistics

IIT Tirupati, Tirupati

September 24, 2024

42