S EC . 7.

C OMPOSITE T RAPEZOIDAL AND S IMPSON S RULE

365

Simpson's Rule

Theorem 7.3 (Composite Simpson Rule). Suppose that [a, b] is subdivided into
2M subintervals [xk , xk+1 ] of equal width h = (b a)/(2M) by using xk = a + kh for
k = 0, 1, . . . , 2M. The composite Simpson rule for 2M subintervals can be expressed
in any of three equivalent ways:
(4a)

M
h
S( f, h) =
( f (x2k2 ) + 4 f (x2k1 ) + f (x2k ))
3
k=1

or
(4b)

S( f, h) =

h
( f0 + 4 f1 + 2 f2 + 4 f3
3
+ + 2 f 2M2 + 4 f 2M1 + f 2M )

366

C HAP. 7

N UMERICAL I NTEGRATION

y
3

y = f(x)

x
0

Figure 7.7 Approximating thearea

or
(4c)

S( f, h) =

M1
M
h
2h 
4h 
( f (a) + f (b)) +
f (x2k ) +
f (x2k1 ).
3
3
3
k=1

k=1

This is an approximation to the integral of f (x) over [a, b], and we write


(5)

f (x) d x S( f, h).

Proof. Apply Simpsons rule over each subinterval [x2k2 , x2k ] (see Figure 7.7). Use
the additive property of the integral for subintervals:


f (x) d x =

(6)

M 


x2k

f (x) d x

k=1 x2k2
M

k=1

h
( f (x2k2 ) + 4 f (x2k1 ) + f (x2k )).
3

Since h/3 is a constant, the distributive law of addition can be applied to obtain (4a). Formula (4b) is the expanded version of (4a). Formula (4c) groups all
the intermediate terms in (4b) that are multiplied by 2 and those that are multiplied
by 4.

Approximating f (x) = 2 + sin(2 x) with piecewise quadratic polynomials produces places where the approximation is close and places where it is not. To achieve
accuracy the composite Simpson rule must be applied with several subintervals. In
the next example we have chosen to integrate this function numerically over [1, 6] and
leave investigation of the integral over [0, 1] as an exercise.

S EC . 7.2

C OMPOSITE T RAPEZOIDAL AND S IMPSON S RULE

367

Example 7.6. Consider f (x) = 2 + sin(2 x). Use the composite Simpson rule with 11
sample points to compute an approximation to the integral of f (x) taken over [1, 6].
To generate 11 sample points, we must use M = 5 and h = (6 1)/10 = 1/2. Using
formula (4c), the computation is
1
1
1
S( f, ) = ( f (1) + f (6)) + ( f (2) + f (3) + f (4) + f (5))
2
6
3
2
+ ( f ( 32 ) + f ( 52 ) + f ( 72 ) + f ( 92 ) + f ( 11
2 ))
3
1
= (2.90929743 + 1.01735756)
6
1
+ (2.30807174 + 1.68305284 + 1.24319750 + 1.02872220)
3
2
+ (2.63815764 + 1.97931647 + 1.43530410 + 1.10831775 + 1.00024140)
3
1
2
1
= (3.92665499) + (6.26304429) + (8.16133735)
6
3
3
= 0.65444250 + 2.08768143 + 5.44089157 = 8.18301550.

Error Analysis
The significance of the next two results is to understand that the error terms E T ( f, h)
and E S ( f, h) for the composite trapezoidal rule and composite Simpson rule are of
the order O(h 2 ) and O(h 4 ), respectively. This shows that the error for Simpsons
rule converges to zero faster than the error for the trapezoidal rule as the step size h
decreases to zero. In cases where the derivatives of f (x) are known, the formulas
E T ( f, h) =

(b a) f (2) (c)h 2
12

and

E S ( f, h) =

(b a) f (4) (c)h 4
180

can be used to estimate the number of subintervals required to achieve a specified

S EC . 7.2

C OMPOSITE T RAPEZOIDAL AND S IMPSON S RULE

369

Corollary 7.3 (Simpsons Rule: Error Analysis). Suppose that [a, b] is subdivided
into 2M subintervals [xk , xk+1 ] of equal width h = (b a)/(2M). The composite
Simpson rule
(14)

S( f, h) =

M1
M
h
2h 
4h 
( f (a) + f (b)) +
f (x2k ) +
f (x2k1 )
3
3
3
k=1

k=1

is an approximation to the integral

(15)

f (x) d x = S( f, h) + E S ( f, h).

Furthermore, if f C 4 [a, b], there exists a value c with a < c < b so that the error
term E S ( f, h) has the form
(b a) f (4) (c)h 4
= O(h 4 ).
180

Example 7.7. Consider f (x) = 2 + sin(2 x). Investigate the error when the composite trapezoidal rule is used over [1, 6] and the number of subintervals is 10, 20, 40, 80,
and 160.
Table 7.2 shows the approximations T ( f, h). The antiderivative of f (x) is

sin(2 x)
,
F(x) = 2x x cos(2 x) +
2

(16)

E S ( f, h) =

and the true value of the definite integral is

x=6

f (x) d x = F(x)
= 8.1834792077.
x=1

This value was used to compute the values E T ( f, h) = 8.1834792077 T ( f, h) in Table 7.2. It is important to observe that when h is reduced by a factor of 12 the successive
errors E T ( f, h) are diminished by approximately 14 . This confirms that the order is O(h 2 ).

Example 7.8. Consider f (x) = 2 + sin(2 x). Investigate the error when the composite
Simpson rule is used over [1, 6] and the number of subintervals is 10, 20, 40, 80, and 160.
Table 7.3 shows the approximations S( f, h). The true value of the integral is
8.1834792077, which was used to compute the values E S ( f, h) = 8.1834792077 S( f, h)
in Table 7.3. It is important to observe that when h is reduced by a factor of 12 , the suc1
cessive errors E S ( f, h) are diminished by approximately 16
. This confirms that the order
4
is O(h ).

370

C HAP. 7

N UMERICAL I NTEGRATION
Table 7.2 Composite
Trapezoidal Rule for

f (x) = 2 + sin(2 x) over [1, 6]

T ( f, h)

10
20
40
80
160

0.5
0.25
0.125
0.0625
0.03125

8.19385457
8.18604926
8.18412019
8.18363936
8.18351924

E T ( f, h) = O(h 2 )
0.01037540
0.00257006
0.00064098
0.00016015
0.00004003

Table 7.3 Composite

f (x) = 2 + sin(2 x) over [1, 6]

S( f, h)

E S ( f, h) = O(h 4 )

5
10
20
40
80

0.5
0.25
0.125
0.0625
0.03125

8.18301549
8.18344750
8.18347717
8.18347908
8.18347920

0.00046371
0.00003171
0.00000204
0.00000013
0.00000001

Example 7.9. Find the number M and the step size h so that the error E T ( f, h) for the
7
composite trapezoidal rule is less than 5 109 for the approximation 2 d x/x T ( f, h).
The integrand is f (x) = 1/x and its first two derivatives are f  (x) = 1/x 2 and
(2)
f (x) = 2/x 3 . The maximum value of | f (2) (x)| taken over [2, 7] occurs at the endpoint
x = 2, and thus we have the bound | f (2) (c)| | f (2) (2)| = 14 , for 2 c 7. This is used
with formula (9) to obtain
(17)

|E T ( f, h)| =

(7 2) 14 h 2
5h 2
| (b a) f (2) (c)h 2 |

=
.
12
12
48

The step size h and number M satisfy the relation h = 5/M, and this is used in (17) to get
the relation
(18)

|E T ( f, h)|

125
5 109 .
48M 2

Now rewrite (18) so that it is easier to solve for M:

25
109 M 2 .
48

Solving (19), we find that 22821.77 M. Since M must be an integer, we choose M =

S EC . 7.2

C OMPOSITE T RAPEZOIDAL AND S IMPSON S RULE

371

possibility that the rounded-off function evaluations will produce a significant amount of
error. When the computation was performed, the result was

5
T f,
= 1.252762969,
22,822
7
which compares favorably with the true value 2 d x/x = ln(x)|x=7
x=2 = 1.252762968. The
error is smaller than predicted because the bound 14 for | f (2) (c)| was used. Experimentation
shows that it takes about 10,001 function evaluations to achieve the desired accuracy of
5 109 , and when the calculation is performed with M = 10,000, the result is

5
= 1.252762973.


T f,
10,000

The composite trapezoidal rule usually requires a large number of function evaluations to achieve an accurate answer. This is contrasted in the next example with
Simpsons rule, which will require significantly fewer evaluations.
Example 7.10. Find the number M and the step size h so that the error E S ( f, h) for the
7
composite Simpson rule is less than 5 109 for the approximation 2 d x/x S( f, h).
The integrand is f (x) = 1/x, and f (4) (x) = 24/x 5 . The maximum value of | f (4) (c)|
taken over [2, 7] occurs at the endpoint x = 2, and thus we have the bound | f (4) (c)|
| f (4) (2)| = 34 for 2 c 7. This is used with formula (16) to obtain
(7 2) 34 h 4
h4
| (b a) f (4) (c)h 4 |

=
.
180
180
48
The step size h and number M satisfy the relation h = 5/(2M), and this is used in (20) to
get the relation
(20)

|E S ( f, h)| =

625
5 109 .
768M 4
Now rewrite (21) so that it is easier to solve for M:
(21)

|E S ( f, h)|

125
109 M 4 .
768
Solving (22), we find that 112.95 M. Since M must be an integer, we chose M = 113,
and the corresponding step size is h = 5/226 = 0.02212389381. When the composite
Simpson rule was performed, the result was

5
= 1.252762969,
S f,
226
7
which agrees with 2 d x/x = ln(x)|x=7
x=2 = 1.252762968. Experimentation shows that it
takes about 129 function evaluations to achieve the desired accuracy of 5 109 , and when
the calculation is performed with M = 64, the result is

5
= 1.252762973.


S f,
128

(22)

372

C HAP. 7

N UMERICAL I NTEGRATION

So we see that the composite Simpson rule using 229 evaluations of f (x) and
the composite trapezoidal rule using 22,823 evaluations of f (x) achieve the same ac1
curacy. In Example 7.10, Simpsons rule required about 100
the number of function
evaluations.

Numerical Methods Using Matlab, 4th Edition, 2004