Introduction to Numerical

Integration
 Definitions
 Upper and Lower Sums
 Trapezoid Method (Newton-Cotes Methods)
 Gauss Quadrature
 Examples

1
 Integration
Indefinite Integrals Definite Integrals

2 1 2 1
x x 1
 x dx  2  c 0 xdx  2 
2
0

Indefinite Integrals of a Definite Integrals are

function are functions numbers.
that differ from each
other by a constant.

2
Fundamental Theorem of Calculus
If f is continuous on an interval [a,b] ,
F is antiderivative of f (i.e., F ' (x)  f(x) )
b
 a
f(x)dx  F(b)  F(a)

x2
There is no antiderivative for : e
b
No closed form solution for :  e dx x2
a
3
The Area Under the Curve
One interpretation of the definite integral is:
Integral = area under the curve

f(x)

b
Area   f(x)dx
a

a b
4
 Upper and Lower Sums
The interval is divided into subintervals.
Partition P  a  x0  x1  x2  ...  xn  b

Define
mi  min  f ( x) : xi  x  xi 1 f(x)
M i  max  f ( x) : xi  x  xi 1

n 1
Lower sum L( f , P )   mi  xi 1  xi 
i 0
n 1
Upper sum U ( f , P )   M i  xi 1  xi 
i 0 x0 x1 x2 x3
a b
5
Upper and Lower Sums

n 1
Lower sum L( f , P )   mi  xi 1  xi 
i 0
n 1
Upper sum U ( f , P )   M i  xi 1  xi  f(x)
i 0

L U
Estimate of the integral 
2
U L
Error 
2
x0 x1 x2 x3
a b
6
Example
1
 0
x 2 dx

 1 2 3 
Partition : P  0, , , ,1
 4 4 4 
n  4 (four equal intervals)
1 1 9
m0  0, m1  , m2  , m3 
16 4 16
1 1 9
M 0  , M1  , M 2  , M 3  1
16 4 16

1 1 1 3
xi 1  xi  for i  0,1, 2, 3 0 1
4 4 2 4

7
Example
n 1
Lower sum L( f , P)   mi xi 1  xi 
i 0

1 1 1 9  14
L( f , P )   0    
4  16 4 16  64
n 1
Upper sum U ( f , P)   M i  xi 1  xi 
i 0

11 1 9  30
U ( f , P)      1 
4 16 4 16  64
1  30 14  11
Estimate of the integral     
2  64 64  32
1  30 14  1
Error     
2  64 64  8 1 1 3
0 1
4 2 4
8
Upper and Lower Sums
• Estimates based on Upper and Lower
Sums are easy to obtain for monotonic
functions (always increasing or always
decreasing).
• For non-monotonic functions, finding
maximum and minimum of the function
can be difficult and other methods can be
more attractive.

9
Newton-Cotes Methods
 In Newton-Cote Methods, the function
is approximated by a polynomial of
order n.
 Computing the integral of a polynomial
is easy.


a
b
f ( x)dx  
a
b
a 0  a1 x  ...  an x dx
n

b (b 2  a 2 ) (b n 1  a n 1 )

a
f ( x)dx a0 (b  a )  a1
2
 ...  an
n 1

10
Newton-Cotes Methods
 Trapezoid Method (First Order Polynomials are used)

b b
a f ( x)dx  
a
a0  a1x dx

 Simpson 1/3 Rule (Second Order Polynomials are used)

a
b
f ( x)dx  
b
a
a0
2

 a1x  a2 x dx

11
 Trapezoid Method
 Derivation-One Interval
 Multiple Application Rule
 Estimating the Error
 Recursive Trapezoid Method

12
 Trapezoid Method
b
I   f ( x)dx
f (b)  f (a) a
f (a)  ( x  a)
ba b  f (b)  f (a) 
I    f (a)  ( x  a) dx
f(x)
a
 ba 
b
 f (b)  f (a ) 
  f (a)  a x
 ba  a
2 b
f (b)  f (a ) x

ba 2 a
a b f (b)  f (a)
 b  a 
2
13
Trapezoid Method
Derivation-One Interval
b 
b f (b)  f (a ) 
I   f ( x)dx    f (a )  ( x  a ) dx
a a
 ba 
b f (b)  f (a ) f (b)  f (a ) 
I    f (a)  a  x dx
a
 ba ba 
b 2 b
 f (b)  f (a )  f (b)  f (a ) x
  f (a)  a x 
 ba  a ba 2 a

 f (b)  f (a )  f (b)  f (a ) 2
  f (a)  a b  a   (b  a 2 )
 ba  2(b  a )
f (b)  f (a )
 b  a 
2
14
Trapezoid Method
f(x)

f (b)

f (a )
ba
Area   f (a)  f (b) 
2
a b
15
 Trapezoid Method
Multiple Application Rule
f ( x2 )  f ( x1 )
Area  x2  x1 
f(x) 2
The interval [a, b] is
partitioned into n segments
a  x0  x1  x2  ...  xn  b
b

a
f ( x)dx  sum of the areas
of the trapezoids
x
x0 x1 x2 x3
a b
16
Trapezoid Method
General Formula and Special Case

If the interval is divided into n segments (not necessarily equal)

a  x0  x1  x2  ...  xn  b
n 1
1
f ( x)dx   xi 1  xi  f ( xi 1 )  f ( xi ) 
b
 a
i 0 2

Special Case ( Equaliy spaced base points)

xi 1  xi  h for all i
1 n 1

f ( x)dx  h   f ( x0 )  f ( xn )   f ( xi ) 
b
 a
2 i 1 
17
Example
Given a tabulated Time (s) 0.0 1.0 2.0 3.0
values of the velocity of
an object. Velocity (m/s) 0.0 10 12 14

Obtain an estimate of
the distance traveled in
the interval [0,3].

Distance = integral of the velocity

3
Distance   0
V ( t ) dt
18
Example 1
The interval is divided Time (s) 0.0 1.0 2.0 3.0

into 3 subintervals Velocity 0.0 10 12 14

Base points are0,1,2,3

(m/s)

Trapezoid Method
h  xi 1  xi  1
 n 1 1 
T  h  f ( xi )   f ( x0 )  f ( xn ) 
 i 1 2 
 1 
Distance  1(10  12)  (0  14)  29
 2 
19
Error in estimating the integral
Theorem
Assumption : f ' ' ( x) is continuous on [a,b]
Equal intervals (width  h)
Theorem : If Trapezoid Method is used to
b
approximate  a
f ( x)dx then
b  a 2 ''
Error   h f ( ) where   [a,b]
12
ba 2
Error  h max f ' ' ( x)
12 x[ a ,b ]

20
Estimating the Error
For Trapezoid Method

How many equally spaced intervals are


needed to compute  0
sin( x)dx
to 5 decimal digit accuracy ?

21
Example

1
sin( x )dx, find h so that error   105
2
0
ba 2
Error  h max f ' ' ( x )
12 x[ a ,b ]
b   ; a  0; f ' ( x )  cos( x ); f ' ' ( x )   sin( x )
 2 1
f ' ' ( x )  1  Error  h   105
12 2
2 6
 h   105  h  0.00437

(b  a ) 
 n   719 intervals
h 0.00437 22
Example
x 1.0 1.5 2.0 2.5 3.0

f(x) 2.1 3.2 3.4 2.8 2.7

3
Use Trapezoid method to compute :  1
f ( x) dx

n 1
1
Trapezoid T ( f , P )    xi 1  xi  f ( xi 1 )  f ( xi ) 
i 0 2

Special Case : h  xi 1  xi for all i,

 n 1 1 
T ( f , P )  h  f ( xi )   f ( x0 )  f ( xn ) 
 i 1 2 
23
Example
x 1.0 1.5 2.0 2.5 3.0

f(x) 2.1 3.2 3.4 2.8 2.7

3 n1 1 
1 f ( x)dx  h  f ( xi )   f ( x0 )  f ( xn ) 
 i 1 2 
 1 
 0.5  3.2  3.4  2.8  2.1  2.7 
 2 
 5.9
24
Recursive Trapezoid Method
Estimate based on one interval :
f(x)

h ba
ba
R (0,0)   f ( a )  f (b) 
2

a ah

25
Recursive Trapezoid Method
Estimate based on 2 intervals :
f(x)

ba
h
2
ba  1 
R (1,0)  f ( a  h )   f ( a )  f ( b ) 
2  2 

1
R(1,0)  R(0,0)  h f (a  h )
2

Based on previous estimate a ah a  2h

Based on new point
26
Recursive Trapezoid Method
f(x)
ba
h
4
ba
R (2,0)   f (a  h)  f (a  2h)  f (a  3h)
4
1 
  f (a )  f (b) 
2 

1
R(2,0)  R(1,0)  h f (a  h)  f (a  3h)
2

a a  2h a  4h
Based on previous estimate
Based on new points
27
Recursive Trapezoid Method
Formulas

ba
R (0,0)   f (a)  f (b)
2

1  2 ( n1) 
R(n,0)  R(n  1,0)  h   f a  (2k  1)h 
2  k 1 
ba
h n
2
28
Recursive Trapezoid Method
ba
h  b  a, R (0,0)   f (a)  f (b)
2
ba 1 1 
h , R (1,0)  R(0,0)  h  f a  (2k  1)h 
2 2  k 1 
ba 1  2 
h 2 , R (2,0)  R(1,0)  h  f a  (2k  1)h 
2 2  k 1 
2 
2
ba 1
h 3 , R(3,0)  R (2,0)  h  f a  (2k  1)h 
2 2  k 1 
..................
2 
( n 1 )
ba 1
h n , R(n,0)  R (n  1,0)  h   f a  (2k  1)h 
2 2  k 1 
29
Example on Recursive Trapezoid
Use Recursive Trapezoid method to estimate :
 /2

 0
sin( x )dx by computing R(3,0) then estimate the error

n h R(n,0)
0 (b-a)=/2 (/4)[sin(0) + sin(/2)]=0.785398
1 (b-a)/2=/4 R(0,0)/2 + (/4) sin(/4) = 0.948059
2 (b-a)/4=/8 R(1,0)/2 + (/8)[sin(/8)+sin(3/8)] = 0.987116
3 (b-a)/8=/16 R(2,0)/2 + (/16)[sin(/16)+sin(3/16)+sin(5/16)+
sin(7/16)] = 0.996785

Estimated Error = |R(3,0) – R(2,0)| = 0.009669

30
Advantages of Recursive Trapezoid
Recursive Trapezoid:
 Gives the same answer as the standard
Trapezoid method.
 Makes use of the available information to
reduce the computation time.
 Useful if the number of iterations is not
known in advance.

31
 Gauss Quadrature
 Motivation
 General integration formula

32
Motivation
Trapezoid Method :
b  n 1 
 1 
f ( x )dx  h  f ( xi )   f ( x0 )  f ( xn ) 
2
a  i 1 
It can be expressed as :
b n

f ( x )dx  ci f ( xi )
a
i 0
 h i  1,2,..., n  1
where ci  
 0.5 h i  0 and n
33
General Integration Formula
b n

f ( x )dx  ci f ( xi )
a
i 0
ci : Weights xi : Nodes
Problem :
How do we select ci and xi so that the formula
gives a good approximation of the integral?

34
Lagrange Interpolation
b b
a
f ( x)dx   Pn ( x)dx
a

where Pn ( x) is a polynomial that interpolates f(x)

at the nodes : x0 , x1 ,..., xn
 n 
   i ( x) f ( xi ) dx
b b b
a
f ( x)dx   Pn ( x)dx  
a a
 i 0 
n
  f ( x)dx   ci f ( xi )
b b
where ci    i ( x) dx
a a
i 0

35
Example
 Determine the Gauss Quadrature Formula of
2

If the nodes are given as (-1, 0 , 1) 

2
f ( x )dx

 Solution: First we need to find l0(x), l1(x), l2(x)

 Then compute:
2 2 2

c0  l0 ( x )dx, c1  l1( x )dx, c2  l2 ( x )dx

2 2 2
36
Solution
( x  x1)( x  x 2) x ( x  1)
l0 ( x )  
( x 0  x1)( x 0  x 2) 2
( x  x 0)( x  x 2)
l1( x )   ( x  1)( x  1)
( x1  x 0)( x1  x 2)
( x  x 0)( x  x1) x ( x  1)
l2 ( x )  
( x 2  x 0)( x 2  x1) 2
2 2 2

  
x ( x  1) 8 4 x ( x  1) 8
c0  dx  , c1   ( x  1)( x  1)dx   , c2  dx 
2 3 3 2 3
2 2 2
2


8 4 8
The Gauss Quadrature Formula for f ( x )dx  f ( 1)  f (0)  f (1)
3 3 3
-2 37
Using the Gauss Quadrature Formula
Case 1 : Let f ( x )  x 2
2 2
2 16
The exact value for f ( x )dx  x dx 
3
-2 -2
8 4 8
The Gauss Quadrature Formula  f ( 1)  f (0)  f (1)
3 3 3
8 2 4 2 8 2 16
 ( 1)  (0)  (1)  , which is the same exact answer
3 3 3 3

38
Using the Gauss Quadrature Formula
Case 2 : Let f ( x )  x 3
2 2

The exact value for f ( x )dx  x 3dx  0

-2 -2
8 4 8
The Gauss Quadrature Formula  f ( 1)  f (0)  f (1)
3 3 3
8 3 4 3 8 3
 ( 1)  (0)  (1)  0, which is the same exact answer
3 3 3

39
Improper Integrals
Methods discussed earlier cannot be used directly to
approximate improper integrals (one of the limits is  or  )
 Use a transformation like the following
1
b

 
a 1 1
f ( x )dx  2
f  dt , (assuming ab  0)
a
1 t t
b
and apply the method on the new function.
 
 1  

 
1 1 1 
Example : 2
dx  2 2
dt
1
x 0
t  1 
t 
   40