Yu - 1

Eva Yu

Mr. Acre

AP Calculus

10 March 2020

Riemann Sums

Integrals are one of the major calculus topics. They consist of finding the area under a

function that could determine another conceptual value. For example, in a velocity-time graph,

the distance could be found from the area of the graph due to the product of velocity and time

being equal to the distance equation, d=v*t. The four integration methods discussed will be

Riemann Sums, Trapezoid Rule, Simpson’s Rule, and the Mean Value Theorem.

There are many different ways to find the integrals in which Riemann Sums is one of

those examples. The formal definition is shown in Figure 1 below.

Paul A. Foerster “Calculus: Concepts and Applications ​Second Edition​” Textbook

Figure 1. The Formal Definition of Riemann Sums

A Riemann sum is a method used to find the area under a function, and it is only an

approximation. This process consists of partitioning, or dividing, the section of integration into

rectangles. Then, the area of each of those rectangles is found before taking the sum of those
 Yu - 2

values in order to approximate the area of the integration section. Another integration method is

the Trapezoid Rule (Figure 2).

Paul A. Foerster “Calculus: Concepts and Applications ​Second Edition​” Textbook

Figure 2. The Formal Process of the Trapezoid Rule

In this process, the section of integration is partitioned into trapezoids before adding the y-values

from each of the increments specified in order to find the integral approximation. The width of

each increment can also be different than the previous one. Next, the Simpson’s Rule is going to

be the third integration method discussed (Figure 3).

Paul A. Foerster “Calculus: Concepts and Applications ​Second Edition​” Textbook

Figure 3. The Formal Process of Simpson’s Rule

The process for this method includes partitioning the integration section into parabolic regions.

The y-values are multiplied with its corresponding coefficient (4 with the second term and the

following alternating values and 2 starting with the third term) before being added up, then

multiplied with ⅓ of the width.

 Yu - 3

All of the three integration methods are used to approximate the integrals, and they all are

divided into shapes. However, these shapes are different for each method. Rectangles are used in

Riemann Sums, trapezoids in the Trapezoid Rule, and parabolic regions in Simpson’s Sule. Also,

in the Riemann Sums and Trapezoid Rule, each width could differ and an increase in the number

of increments results in a closer approximation to the actual value. The process of Simpson’s

Rule is the most accurate because the parabolic regions resemble the graph more compared to the

spacing left from using rectangles and trapezoids.

Now that the three methods are introduced, it is time to do an example:

● Given f(x) = (x-3)​4​+2(x-3)​3​-4(x-3)+5 on the interval from x=1 to x=5, illustrate the

following 5 Riemann Sums with 2 intervals as well as finding their sums: left, right,

midpoint, upper, and lower.

These solutions will be shown in the following figures.

Figure 4. Left Riemann Sum Graph

 Yu - 4

Figure 4 displays the illustration of a left Riemann Sum where the sample points are

taken from the left endpoint of each subinterval, shown in red dots. In this case, the left endpoint

of each subinterval is at x=1 and x=3. The height of each rectangle can be found by finding the

y-value of the sample point. This means that the height for the first rectangle would be 13

because the y-value of the sample point at x=1 is 13. As for the second rectangle, the height

would be 5 since the y-value of the sample point at x=3 is 5. The calculation to find the sum of

the two rectangles is shown below:

Left​n​ = △x (f(c​1​) + f(c​2​)) Only two sample points are used since there
= f(c​1​)( △x ) + f(c​2​)( △x ) are only two intervals

Left​2​ = 2 (f(1) + f(3)) The sample points taken from the left
endpoints are at x=1 and x=3. Also, the width
= f(1)(2) + f(3)(2) for both subintervals is the same.

Left​2​ = 2 (13 + 5) Simplify

= 26 + 10

= 36 u​2 Evaluate

The sum of the two rectangles for the left Riemann Sum is found to be 36. In the table, two

versions of the formula are used. The first one is used when the width of each subinterval is the

same while the second one is used for both when the widths are either the same or different. Due

to the width being the same in this situation, both versions of the formula resulted in the same

value.
 Yu - 5

Note: The function extends beyond the graph shown here, so the second rectangle would be
much taller since f(5)=29.
Figure 5. Right Riemann Sum Graph

Figure 5 illustrates the right Riemann Sum and the sample points are taken from the right

endpoints of each subinterval. From the interval from x=1 to x=5, the right endpoints are at x=3

and x=5 with each height being 5 and 29 respectively. The calculation to find the sum of the two

rectangles is shown below:

Right​n​ = △x (f(c​1​) + f(c​2​)) Only two sample points are used since there
= f(c​1​)( △x ) + f(c​2​)( △x ) are only two intervals

Right​2​ = 2 (f(3) + f(5)) The sample points taken from the right
endpoints are at x=3 and x=5. Also, the width
= f(3)(2) + f(5)(2) for both subintervals is the same.

Right​2​ = 2 (5 + 29) Simplify

= 10 + 58

= 68 u​2 Evaluate
 Yu - 6

The sum of the two rectangles for the right Riemann Sums is found to be 68.

Figure 6. Midpoint Riemann Sum Graph

The illustration for the midpoint Riemann Sum is shown in Figure 6. The sample points

are taken from the midpoint of each subinterval in which they are at x=2 and x=4 with a height

of 8 and 4 respectively. The calculation to find the sum of the two rectangles is shown below:

M​n​ = △x (f(c​1​) + f(c​2​)) Only two sample points are used since there
= f(c​1​)( △x ) + f(c​2​)( △x ) are only two intervals

M​2​ = 2 (f(2) + f(4)) The sample points taken from the midpoint of
each subinterval are at x=2 and x=4. Also, the
= f(2)(2) + f(4)(2) width for both subintervals is the same.

M​2​ = 2 (8 + 4) Simplify

= 16 + 8

= 24 u​2 Evaluate
 Yu - 7

The sum of the two rectangles for the midpoint Riemann Sums is found to be 24.

Note: The function extends beyond the graph shown here, so the second rectangle would be
much taller since f(5)=29.
Figure 7. Upper Riemann Sum Graph

The Upper Riemann Sum is shown in Figure 7 where there is an overestimation. The

sample points are taken where the f(x) value is the highest at each subinterval. For this situation,

the highest f(x) value in the first subinterval is at x=1 with a height of 13. As for the second

subinterval, the highest f(x) value is at x=5 with a height of 29. The calculation to find the sum

of the two rectangles is shown below:

U​n​ = △x (f(c​1​) + f(c​2​)) Only two sample points are used since there
= f(c​1​)( △x ) + f(c​2​)( △x ) are only two intervals

U​2​ = 2 (f(1) + f(5)) The sample points taken from the highest f(x)
value of each subinterval are at x=1 and x=5.
= f(1)(2) + f(5)(2) Also, the width for both subintervals is the
same.

U​2​ = 2 (13 + 29) Simplify

= 26 + 58

= 84 u​2 Evaluate
 Yu - 8

The sum of the two rectangles for the midpoint Riemann Sums is found to be 84.

Figure 8. Lower Riemann Sum Graph

Figure 8 shows the illustration of the lower Riemann Sum in which there is an

underestimation. The sample points are taken from the lowest f(x) value at each subinterval and

there were found to be at x=3 and x=3.68 with a height of 5 and 3.12 respectively. The

calculation to find the sum of the two rectangles is shown below:

L​n​ = △x (f(c​1​) + f(c​2​)) Only two sample points are used since there
= f(c​1​)( △x ) + f(c​2​)( △x ) are only two intervals

L​2​ = 2 (f(3) + f(3.68)) The sample points taken from the lowest f(x)
value of each subinterval are at x=3 and
= f(3)(2) + f(3.68)(2) x=3.68. Also, the width for both subintervals
is the same.

L​2​ = 2 (5 + 3.12) Simplify

= 10 + 6.24

= 16.24 u​2 Evaluate

 Yu - 9

The sum of the two rectangles for the midpoint Riemann Sums is found to be 16.24.

The next example is going to consist of the other two integration methods.

● Given f(x) = (x-3)​4​+2(x-3)​3​-4(x-3)+5 on the interval from x=1 to x=5, illustrate the

Trapezoid Rule and Simpson’s Rule with 4 intervals as well as finding their sums.

These solutions will be shown in the following figures.

Note: The function extends beyond the graph shown here, so the second rectangle would be
much taller since f(5)=29.
Figure 9. Trapezoid Rule Graph

The graph of the Trapezoid Rule is shown above in Figure 9. Since the integration section is

partitioned into trapezoids, the formula for finding a trapezoid can be used along with the f(x)

values representing the base values. The calculation for the sum of the four trapezoids is shown

below:

T​n​ = 0.5(f(x​1​)+ f(x​2​))( △x ) + 0.5(f(x​2​)+ f(x​3​))( △x ) Sum of all the areas of the four
+ 0.5(f(x​3​)+ f(x​4​))( △x ) + 0.5(f(x​4​)+ f(x​5​))( △x ) trapezoids
 Yu - 10

T​n​ = ( △x )(0.5f(a) + f(x​1​) + f(x​2​) + f(x​3​) + 0.5f(b)) f(a) and f(b) are the first and last values
respectively

T​4​ = 0.5(f(1) + f(2))( 1 ) + 0.5(f(2) + f(3))(1) + Substitute values

0.5(f(3) +f(4))( 1 ) + 0.5(f(4) + f(5))( 1 )

T​n​ = ( 1 )(0.5f(1) + f(2) + f(3) + f(4) + 0.5f(5))

T​4​ = 0.5(13+8)(1) + 0.5(8+5)(1) + 0.5(5+4)(1) + Input values

0.5(4+29)(1)

T​n​ = (1)(0.5(13) + 8 + 5 + 4 + 0.5(29))

= 38 u​2 Evaluate

The sum of areas of the four trapezoids is found to be 38. There are two versions of the

Trapezoid Rule formula shown in the table above with the first formula providing a better

conceptual understanding of adding the trapezoid areas together. As for the second formula, it is

more time-efficient.
 Yu - 11

Note: The function extends beyond the graph shown here, so the second rectangle would be
much taller since f(5)=29.
Figure 10. Simpson’s Rule Graph

The Simpson’s Rule graph is illustrated above and the formula for finding the area of a parabola

will be used because the integration section is partitioned into parabolic regions. When the

function is concave up in the subinterval, there is an overestimation. The calculation for the sum

of the parabolic regions is shown below:

I = ⅓ ( △x )(y​0​+4y​1​+2y​2​+4y​3​+y​n​) Simpson’s Rule Property

I = ⅓ ( 1 )(f(1)+4f(2)+2f(3)+4f(4)+f(5)) Substitute values

I = ⅓ ( 1 )(13+4(8)+2(5)+4(4)+29) Input values

= 33.33 u​2 Evaluate

The sum of the areas of the parabolic regions is found to be 33.33.

After computing the equations from the three methods, it is time to make a comparison.

The definite integral of the function, f(x) = (x-3)​4​+2(x-3)​3​-4(x-3)+5, on the interval from x=1 to

x=5 is shown in Figure 11 below.

b
Definite integral
∫ f(x) dx
a

5
Input values
∫ (x-3)​4​+2(x-3)​3​-4(x-3)+5 dx
1

= 32.8 u​2 Evaluate

Figure 11. Definite Integral of f(x)

All of the approximations calculated from the methods are similar in value with the exception of

four of the Riemann Sums being much greater or less than the exact integration value. The value

calculated from using Trapezoid Rule is close to the definite integral but the left Riemann Sum is
 Yu - 12

still closer to the actual integral value. However, out of all the methods, the Simpson’s Rule is

the most accurate with a difference of only 0.533. As mentioned before, the Simpson’s Rule

process is considered the most accurate because the parabolic regions resemble the shape of the

function the closest.

Another integration method is the Mean Value Theorem. In this method, the area is found

when the average value of the integral is multiplied with the interval, or its base, much like the

area formula for a rectangle.

Figure 13. Mean Value Theorem Graph

Figure 13 illustrates the integration graph when using the Mean Value Theorem. As shown, the

widths are the same along with the heights in which they intersect with the function. The area

under the curve is found in Figure 14 below using two intervals.

 Yu - 13

b
Definite integral
∫ f(x) dx
a

3 5
Input values
∫ (x-3)​4​+2(x-3)​3​-4(x-3)+5 dx + ∫ (x-3)​4​+2(x-3)​3​-4(x-3)+5 dx
1 3

= 32.8 u​2 Evaluate

Figure 14. Integration using Two Intervals

Though the integration was split into two intervals, it still results in the same value as when one

interval is used (Figure 11).

b b
Height = 1
b−a ∫ f(x) dx Height = 1
b−a ∫ f(x) dx
a a

3 5
1
3−1 ∫ (x-3)​4​+2(x-3)​3​-4(x-3)+5 dx 1
5−3 ∫ (x-3)​4​+2(x-3)​3​-4(x-3)+5 dx
1 3

= 8.2 u = 8.2 u

Figure 15. Finding the Height of Each Rectangle

The height of both rectangles can be found by dividing the value of the integral by the interval of

“b-a” (Figure 15). The process for finding the height of the first rectangle is shown on the left

while the second rectangle is shown on the right. The area of these rectangles can be found by

multiplying their height with its base. These two areas are then added together to find the integral

for the overall interval. This calculation is shown below in Figure 16:

b b
∫ f(x) dx = f(c)(b-a) ∫ f(x) dx = f(c)(b-a)
a a

= 8.2(3-1) = 8.2(5-3)
 Yu - 14

= 16.4 u​2 = 16.4 u​2

Figure 16. Finding the Area Using the Mean Value Theorem

As mentioned before, the area under the function can be found using the Mean Value Theorem

by adding the product of the two rectangles’ heights with the base. In this situation, both of the

rectangles’ areas are 16.4 and adding them together results in 32.8, which is the value of the

definite integral for when the interval is from x=1 to x=5.

It is time for another problem. This one is much more interesting:

● The volume of a spherical hot air balloon expands as the air inside the balloon is heated.

The radius of the balloon, in fact, is modeled by a twice-differentiable function r of time

t, where t is measured in seconds. For 0<t<12, the graph is concave down. The table

below gives selected values of the rate of change, r’(t), of the radius if the balloon over

the time interval 0<t<12. The radius of the balloon is 32 feet when t=7. (The volume of a

sphere of radius r is given by V=4/3 π r​3​.

t (seconds) 0 1 4 7 11 12

r’(t) (ft/sec) 5.7 4.0 2.0 1.4 0.5 0.4

a) Estimate the radius of the balloon when t=7.2 using the tangent line

approximation at t=7. Is your estimate greater than or less than the true value?

Y = 1.4(t-7)+32 ← tangent line equation

The slope of the tangent line is the derivate in which it displayed in the

table as 1.4 when x=7, and since the x-value of 7 is used, the y-value of 32

is used for the translation. After substituting in 7.2 for t, the estimation is
 Yu - 15

found to be 32.28. It is an overestimation because the graph is concave

down and the tangent line is on the outside of the graph.

b) Find the rate of change of the volume of the balloon with respect to the time when

t=7. Indicate the units of measure.

To do this part of the problem, the volume equation must be differentiated:

V=4/3 π r​3

V’ = 4 π r​2 ​* r’

V’ = 4 π ( 32)​2 ​* 1.4 = 5734 π ft​3​/s

c) Use a right Riemann Sum with 5 subintervals indicated by the data in the table to
12
approximate ∫ r′(t) dt while also explaining the meaning of the integral in terms
0

of the balloon’s radius.

f(c​1​)( △x ) + f(c​2​)( △x )+f(c​3​) + ( △x ) + f(c​4​)( △x ) + f(c​5​)( △x )

= 4(1-0) + 2(4-1) + 1.4(7-4) + 0.5(11-7) + 0.4(12-11)

= 16.6 ft

This means that the radius is increasing at a rate of 16.6 feet from 0 to 12

seconds.
12
d) Is your approximation in part c greater than or less than ∫ r′(t) dt ?
0

The right Riemann Sum is an overestimation of the value since it uses the

highest f(x) value in each subinterval considering the graph is concave

down.
 Yu - 16

There are many methods to finding the integral of a function. Different shapes such as

rectangles, trapezoids, and parabolic regions may be used to find the approximation of those

integrations. When approximating, the Simpson’s Rule is the most accurate. These methods are

useful when there are data points provided and there is no function or there is a function, but the

fundamental theorem of calculus is difficult to solve.