You are on page 1of 6

Roberto’s Notes on Integral Calculus

Chapter 6: Improper and approximate integration Section 5

Approximate integration:
The midpoint method
What you need to know already: What you can learn here:
 What a definite integral is and how to  The midpoint method to approximate the
compute it through the FTC. value of a definite integral.

The FTC allows us to compute definite integrals in a fairly easy way, but only
provided we know how to obtain the antiderivatives of the integrand. However, for b n
many functions computing such antiderivative is very difficult and sometimes even
impossible. What can we do in such cases? The next best approach is to find a way a f ( x)dx  
i =1
f ( xi ) x
to approximate the value, especially if the method we come up with can be taken to
any degree of accuracy that we need.
For any choice of n and of the values of xi in each
We have already seen the simplest such method, although we saw it from a interval we can obtain a different approximation and
different perspective, namely within the context of building up the concept of we can adapt such choices in order to get better
definite integral.
approximations.

Strategy for
approximating definite integrals: It turns out that there is a wise and simple choice for the representative values
xi that works fairly well in a wide variety of situations. This is obtained by picking
The Riemann sum rule the midpoint of each interval:
b n
Since  f ( x)dx = lim  f ( x ) x , we can
a
n→
i =1
i

approximate the value of a definite integral by using


the Riemann sum formula:

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 1

Strategy for
approximating definite integrals:
Example:
 sin(sin(x))dx
0

The midpoint rule Finding an antiderivative for this integrand is not possible, but we can use the
midpoint rule to approximate the value of this definite integral.

If f ( x ) is a continuous function on a, b , a good  
b For the sake of illustration, let us use n=6,. In that case x = , so that the
6
way to approximate the definite integral  f ( x)dx is intervals we use are:
a              2   2 5   5 
to use the Riemann sum approximation with intervals 0, 6  ,  6 , 3  ,  3 , 2  ,  2 , 3  ,  3 , 6  ,  6 ,  
of equal length and to select the midpoint of each
interval as representative value for the function. and their midpoints are, respectively:

In this way we let: xi =


( 2i − 1)  :  3 5 7 9 11
,, , , ,
b−a 12 12 12 12 12 12 12
x = Therefore the approximate value of the integral is:
n
 −0      3   5 
xi = a + ix −
x
=a+
( 2i − 1) x M= sin  sin  + sin  sin  + sin  sin  +
6   12   12   12 
2 2
 7   9   11  
+ sin  sin  + sin  sin  + sin  sin 
so that the formula becomes:
b
 12   12   12  

 f ( x)dx 
a
This quantity can now be computed with the help of a friendly calculator and
turns out to be approximately 1.81.

b−a   x   3x   x 
 f a + + f a +  + ... + f  b − 
n   2   2   2  And speaking of computing machines, notice that it is fairly easy to translate
the midpoint rule into a computer program, given its repetitive structure. In fact, if
we want a good approximation with this method, we may prefer a large value of n.
Such a large value presents no difficulty to a computer, but can put your mental
sanity at risk if you attempt to do it by hand!

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 2
Y
10
Knot on your finger 9
8
7
Although the midpoint rule provides only an 6
approximation of the required integral, if the integral 5
exists, such as when the integrand is continuous, we 4
3
can make such approximation as accurate as we
2
want by choosing a large enough number of slices. 1 X
0 1 2 3 4 5 6 7 8 9 10

This gives us an idea for how to possibly improve on this method: change the
Notice also that in this method we approximate each slice by using a rectangle. way we top each slice by using a more advanced type of functions that may
This means that each slice of the region is approximated by using, as top boundary, approximate the given curve better.
not the original function, but a constant function of the form y = f ( xi ) . We shall try to use this idea in the next two sections, although I warn you not to
expect too much: the improvements we shall achieve turn out to be disappointing at
first, but the exploration and the information they provide are worth the effort and
will reap good fruits, but in later courses!

Summary
 A finite Riemann sum may be used as a reasonable approximation for a definite integral when the FTC is difficult or impossible to apply.
 By using a Riemann sum with slices of equal width and by choosing the midpoint of each slice as the value at which to compute the function, we obtain the midpoint rule, a
very efficient method that can be made as accurate as desired by increasing the number of slices used.

Common errors to avoid


 Spend some time to identify the width x and all the midpoints to be used. Students often rush through this step and end up using incorrect values leading to incorrect
approximations.

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 3
Learning questions for Section I 6-5

Review questions:

1. Explain when an approximate integration method may be useful. 3. Explain why the midpoint rule is ideal for computer implementation.

2. Describe how the formula for the midpoint rule is constructed.

Memory questions:

1. Which is the formula for the midpoint rule?

Computation questions:
In each of questions 1-4, compute the midpoint estimate of the given integral by using the given number of intervals. Then compare it to the exact value, obtained by applying the
FTC.

4 2
1.  ln
1
x dx, n = 6 3.  cosh x dx,
0
n=5

3 4


 xe dx, n = 6  1/ x
2
x
2. 4. A = dx, n = 6
0 x2
1

In each of questions 5-8, compute the midpoint estimate of the given interval by using the given number of intervals. Then compare it to the value, obtained by using a computer
or calculator program.

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 4
9
6
7. A =
 5 − x dx, n = 6, A  9.61
5. A =
 sin
0
x dx, n = 6, A  5.048 3


1
2 8. A = dx, n = 6, A  2.59
6. A =  cosh x dx, n = 5, A  8.67
2
ln x
2
0

9. Set up the integral that represents the length of the curve 11. Use the midpoint rule to estimate the value of ln 7 . To do that, remember that
y = x + ln x, 1  x  4 and then the formula that represents its midpoint 1
approximation for n=6. No need to actually compute the formula.
y = ln x is an antiderivative for y = , and use n=6.
x
10. Use the midpoint rule to approximate the area of a quarter circle of radius 2 with 12. Use the same method as in the previous question to estimate the value of ln 5
n=4 and compare it to the exact value.
with n=4.

The integrals of questions 12-13 are both improper and convergent. Compute their value and then compare it to the value obtained by using a midpoint estimate with n=6.

3 3

 
1 1
13. dx 14. dx
x 3− x
0 −1

Theory questions:

1. What type of curves are used to approximate the top boundary of each strip in 3. What can be done to improve an approximations obtained by using the midpoint
the midpoint rule? rule?

2. Why is a left Riemann sum NOT a good approximation method in general?

Application questions:

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 5
a
1. Use the midpoint rule to estimate the value of  e− x dx for a value of a of your choice. Notice that this integral cannot be computed exactly, but it is an important one, as it is
2

0
related to the probability values of a normal (Gaussian, bell-shaped) distribution.

Templated questions:

1. Construct a simple definite integral and use the midpoint rule to estimate its value.

What questions do you have for your instructor?

Integral Calculus Chapter 6: Improper and approximate integration Section 5: The midpoint method Page 6

You might also like