Chapter 4: Integrals

Part 3

By
Assoc.Prof. Mai Duc Thanh
 Outline
1. Approximate Integration
1. Mid-Point Rule
2. Trapezoidal Rule
3. Simpson’s Rule
2. Improper Integrals
1. Improper Integrals of Type 1
2. Improper Integrals of Type 2

9/11/2016 2
 1. Approximate Integration:
Motivations
• Integration techniques can be applied to only certain
classes of (simple) functions
Many integrals cannot be evaluated by integration
techniques. Examples

1 2 2
sin x
0 1 x 1 dx,...
2
x x
e dx, dx , x

• In many applications: Functions are in tabular form

• In both cases we need to find approximate values of
definite integrals.
9/11/2016 3
 Mid-Point Rule
-Left endpoint Rule: xi 1  a  (i  1) x
ba
b n

 f ( x)dx  f  x x,
a i 1
i 1 x 
n

-Right endpoint Rule: xi  a  ix

ba
b n

 f ( x)dx  f  x x,
a i 1
i x 
n

 f ( x)dx  f  x x
b n
-Mid-point Rule: i
a i 1

xi 1  xi ba
xi  , x 
2 n
 Trapezoidal Rule
Take average of left-end and right-end point rules gives us
1 n 
b n

 f ( x)dx    f  xi 1 x   f  xi x 

a
2  i 1 i 1 
x  n 
    i 1   i  
2  i 1
f x  f x

x
  f  x0   2 f  x1    2 f  xn 1   f  xn  
2

Trapezoidal Rule is:

x
 a f ( x)dx  Tn  2  f (a)  2  f  x1    f  xn 1    f (b) 
b

ba
x  , xi  a  ix
9/11/2016
n 5
Example 1: Use the Trapezoidal Rule with n=5 to
approximate the integral
2
1
1 x dx
Solution

n  5, a  1, b  2, and x  (2  1) / 5  0.2
2
1 0.2
1 x dx  T5  2  f (1)  2 f (1.2)  2 f (1.4)  2 f (1.6)  2 f (1.8)  f (2)
1 2 2 2 2 1
 0.1        0.695635
 1 1.2 1.4 1.6 1.8 2 
9/11/2016 6
 Example 2: Use the Midpoint Rule with n=5 to
approximate the integral
2
1
1 x dx
Solution

The midpoints of the five subintervals are

1.1, 1.3, 1.5, 1.7, and 1.9, so the Midpoint Rule gives
2
1
1 x dx  M 5  x  f (1.1)  f (1.3)  f (1.5)  f (1.7)  f (1.9)
 1 1 1 1 1 
 0.2        0.691908
 1.1 1.3 1.5 1.7 1.9 
9/11/2016 7
 Errors in Approximating Integrals
In last Examples:
2
1
1 x dx  ln x 1  ln 2  0.693147 ...
2

Errors: ET  0.002488 and EM  0.001239

In general, errors are defined by:

b b
ET   f ( x)dx  Tn and EM   f ( x)dx  M n
a a

9/11/2016 8
 Errors

9/11/2016 9
 Error Bounds
Suppose |f ''( x) | K , a  x  b. If ET and EM are errors
in the Trapezoidal and Midpoint Rules, then
K (b  a) 3
K (b  a ) 3
ET  and EM 
12n 2 24n 2

9/11/2016 10
 K (b  a)3
|f ''( x) | K , a  x  b. It holds that ET 
12n2

Example 1: How large should we take n in order to guarantee

2
that the Trapezoidal Rule approximation for 1
is accurate to within 0.0001? x
dx 
1

Solution: 1 1 2
f ( x)  , f '( x)  2 , f "( x)  3 | f "( x) | 2 for 1  x  2
x x x
Error Bound for Trapezoidal Rule gives:

2(1)3 2 1
| ET | 2
 0.0001  n 
2
n  40.8
12n 12(0.0001) 0.0006
9/11/2016
n  41 will ensure the desired accuracy. 11
 K (b  a)3
|f ''( x) | K , a  x  b. It holds that EM 
24n2

Example 2: How large should we take n in order to guarantee

2
that the Midpoint Rule approximation for 1
is accurate to within 0.0001? x
dx 
1

Solution: 1 1 2
f ( x)  , f '( x)  2 , f "( x)  3 | f "( x) | 2 for 1  x  2
x x x
Error Bound for Midpoint Rule gives:

2(1)3 1
| EM | 2
 0.0001  n   28.8675
24n 0.0012
9/11/2016
 n  29 will ensure the desired accuracy. 12
 Simpson’s Rule

ba
x0  a  x1  x2  ...  xn  b, xi  a  ih, h 
n
i  0,1, 2,..., n, n : even
b
h
 f ( x)dx  Sn   f (a)  4  f ( x1 )  f ( x3 )  ...  f ( xn 1 ) 
a
3
 2  f ( x2 )  f ( x4 )  ...  f ( xn 2 )   f (b) 
9/11/2016 13
Example: Use Simpson’s Rule with n=10 to approximate
2
the integral 1
Solution
 x dx
1
Putting f ( x)  1/ x, n  10, x  0.1 in Simpson's Rule gives us
x
2
1
1 x dx  S10  3  f (1)  4  f (1.1)  f (1.3)  f (1.5)  f (1.7)  f (1.9) 
 2  f (1.2)  f (1.4)  f (1.6)  f (1.8)   f (2) 
0.1 1  1 1 1 1 1 
   4     
3 1  1.1 1.3 1.5 1.7 1.9 
 1 1 1 1  
 2      f (2)   0.693150
 1.2 1.4 1.6 1.8  
Error = Es  0.693147 - 0.693150  0.000003
9/11/2016 14
 Error Bounds

Suppose |f (4) ( x) | K , a  x  b. If ES is the error

in Simpson's Rule, then
K (b  a)5
ES 
180n 4

9/11/2016 15
 K (b  a ) 5
Suppose |f ( x) | K , a  x  b. Then ES 
(4)

180n4
Example: How large should we take n in2order to guarantee
that Simpson’s Rule approximation for 1 dx
is accurate to within 0.0001? 1
x 

Solution: 1 (4) 24
f ( x)  , f ( x)  5 | f (4) ( x) | 24 for 1  x  2
x x
Error Bound for Simpson’s Rule gives:
24(1)5 24 1
| ES | 4
 0.0001  n 
4
n 4  6.04
180n 180(0.0001) 0.00075
 n  8 (n must be even) gives the desired accuracy.
9/11/2016 16
 2. Improper Integrals

•Improper Integrals of Type 1

•Improper Integrals of Type 2

9/11/2016 17
 Definition of an Improper Integral of Type 1
t
a) If 
a
f ( x)dx exists for every number t ≥ a, then
 t
 f ( x)dx  lim  f ( x)dx
a
t 
a

provided this limit exists (as a finite number).

b
b) If t f ( x)dx exists for every number t ≤ b, then
b b


f ( x)dx  lim  f ( x)dx
t 
t

provided this limit exists (as a finite number).

 b
The improper integrals  f ( x)dx and  f ( x)dx are called
a 
convergent if the corresponding limit exists and divergent
if the limit does not exist.
a 
c) If both  f ( x)dx and  f ( x)dx are convergent, then we define
a
 a 


f ( x)dx   f ( x)dx   f ( x)dx
 a
 Examples
t
 1 1 t  1  1 1
1. 1 x 2
dx  lim  2 dx  lim    lim     1
t 
1 x
t   x 1 t   t 1

2. 
0


x
0
e dx  lim  e dx  lim e
t
x
   lim e
x 0
t
0

 et  1
t  t  t 


1 0 1  1
3.  dx   dx   dx
 1  x 2  1  x 2 0 1 x2

 1

 lim tan x t  lim tan x 0
0
 1
 t

t   t 

  

 lim  tan t  lim 1
  
tan t       
1

t   t   2 2
All three integrals are convergent.
An example of a divergent integral:

dx  lim  dx  lim ln x 1  lim ln t  ln 1  

 1 t 1

t
1 x t 
1 x
t  t 

 1
Recall from the previous slide that 1 x 2
dx is convergent

The general rule is the following:

 1
1 x p
dx is convergent if p  1 and divergent if p  1
 Definition of an Improper Integral of Type 2
a) If f is continuous on [a, b) and is discontinuous at b, then
b t
 a
f ( x)dx  lim  f ( x)dx
t b 
a

if this limit exists (as a finite number).

b) If f is continuous on (a, b] and is discontinuous at a, then
b b

a
f ( x)dx  lim  f ( x)dx
t a 
t

if this limit exists (as a finite number).

b
The improper integral f ( x)dx is called convergent if the
a
corresponding limit exists and divergent if the limit does
not exist.
c) If f has a discontinuity at c, where a < c < b, and both
c
 f ( x)dx
b
a
and

c
f ( x)dx are convergent, then we define
b c b

a
f ( x)dx   f ( x)dx   f ( x)dx
a c
Example 1:


1

0
2 x
1
dx  lim 

1

t
1
2 x
dx  lim

 x   lim  1  t   1
1
t
t 0 t 0 t 0 

This integral is convergent.

Example 2:

dx
1
0 x 4

1 1
lim 
2

dx 1
b 0 b x
-1 0 1
1
lim ln x b -1

b 0

lim ln1  ln b  This integral is divergent.

b0
Comparison Theorem for Improper Integrals of Type 1:
Let 0  f ( x)  g ( x), x  a. Then,
 
a)  g ( x)dx is convergent, then  f ( x)dx
a a
is convergent

 
b)  f ( x)dx is divergent, then  g ( x)dx
a a
is divergent

b 

Similar result holds for 


f ( x)dx and 

f ( x)dx

and for Improper Integral of Type 2

9/11/2016 23