Integration

This notebook will show you how to use Mathematica to explore topics introduced in Chapters 5 through 7 of the text.

Computing Sums
The Mathematica command Sum#aj , j, 1, n' is used to compute nj 1 a j . For example, the following input
instructs Mathematica to compute nj 1 +5 j2  4/.
In[1]:=
Out[1]=

Sum#5j2  4, j, 1, 10'

1885

The sum can be entered in a more natural way using ff

e in the palette BasicInput.

10
In[2]:=

+5 j2  4/
j 1

Out[2]=

1885

Riemann Sums
Since Mathematica does not contain a built-in function for illustrating and computing Riemann sums (discussed in Section
5.2 of the text), you can execute the following input cell to create a function which will compute Rn . Don't worry about
understanding the code contained here since it involves a couple of advanced concepts which have not been introduced in
this manual.

Integration

52

In[3]:=

r#a_, b_, n_, m_' : Module$'x

ba
cccccccccccc , recs, est, gr, x, c ,
n

x#i_' : a  i 'x;
c#i_' : x#i  1  m';
n

est

f#c#i'''x;
i 1

recs ListPlot#
Flatten#Table#
x#i  1', 0, x#i  1', f#c#i'',
x#i', f#c#i'', x#i', 0, i, 1, n', 1',
PlotJoined True, DisplayFunction Identity,
PlotRange All, PlotStyle RGBColor#0, 0, 1'';
gr Plot#f#x', x, a, b, DisplayFunction Identity, PlotRange All';
Show#recs, gr, DisplayFunction $DisplayFunction';
Return#est ss N'(
To compute a Riemann sum Rn , define the function f which you are using first and then enter r#a, b, n, 0'if ci xi1 ,
enter r#a, b, n, 1'if ci xi or enter r#a, b, n, .5' if ci cccc12 +xi1  xi /. For example, R6 is computed below for
2
the integral 1 +x3  1/ x using ci xi , ci xi1 and ci cccc12 +xi1  xi /, respectively.
In[4]:=

f#x_' :
r#1, 2,
r#1, 2,
r#1, 2,

x3  1;
6, 0'
6, 1'
6, .5'

1.2
Out[5]=

4.1875

1.4

1.6

1.8

Integration

Out[6]=

1.2

1.4

1.6

1.8

1.2

1.4

1.6

1.8

5.35417

Out[7]=

4.73958

To compute the Riemann sums without the graphical illustration of the rectangles, first execute the contents of the following input cell.
In[8]:=

Clear#r';
ba
cccccccccccc , recs, est, gr, x, c ,
n

r#a_, b_, n_, m_' : Module$'x

x#i_' : a  i 'x;
c#i_' : x#i  1  m';
n

est

f#c#i'''x;
i 1

Return#est ss N'(
2

In the following cell, Rn is computed for the integral 0 #x+x  1/'dx for values of n
ci cccc12 +xi1  xi /.

10, 20, ..., 100 where

53

Integration

54

In[10]:=

Clear#f';
f#x_' x+x  1/;
Table#r#0, 2, n, .5', n, 10, 100, 10' ss TableForm

Out[12]//TableForm=

0.66
0.665
0.665926
0.66625
0.6664
0.666481
0.666531
0.666563
0.666584
0.6666
2

From the output above, it appears that 0 #x+x  1/'dx 0.666.

Indefinite Integrals
Enter Integrate#f #x', x' to evaluate f #x' x.
In[13]:=
Out[13]=

Integrate#3x2  2x  5, x'

5 x  x2  x3

In the above output, note the Mathematica does not include a general constant term C and therefore you must remember
that +3x2  2x  5/dx x3  x2  5x  C.
An indefinite integral can also be evaluated using the button e f found on the palette BasicInput.
In[14]:=

Out[14]=

2
+3x  2x  5/ x

5 x  x2  x3

Definite Integrals
b

Enter Integrate#f #x', x, a, b' to compute a f +x/dx.

In[15]:=
Out[15]=

Integrate#3x2  2x  5, x, 1, 2'

9
f

On the palette BasicInput, click on f e f to enter the integral in a more natural form.
2
In[16]:=

2
+3x  2x  5/ x
1

Out[16]=

Integration

55

Partial Fractions
The Apart command will form the standard decomposition of a rational function.
In[17]:=

Out[17]=

1
cccccccccccccc (
Apart$ cccccccccccccccccccccccccccccccc
+2x  5/+x2  1/x
1
8
2  5 x
cccccccc  cccccccccccccccccccccccccccccccc  cccccccccccccccccccccccc
ccc
5x
145 +5  2 x/
29 +1  x2 /

Before looking at the following output, can you guess that value of the integral based upon the partial fractions in the output
above?
In[18]:=

Out[18]=

% x
5
2 ArcTan#x'
Log#x'
4
 ccccccccccccccccccccccccccccccc  cccccccccccccccccc  cccccccccc Log#5  2 x'  ccccccc Log#1  x2 '
58
29
5
145

Mathematica can be used to compute the integral directly as shown below.

In[19]:=

Out[19]=

1
cccccccccccccc  x
cccccccccccccccccccccccccccccccc
+2x  5/+x2  1/x
2 ArcTan#x'
Log#x'
4
5
 ccccccccccccccccccccccccccccccc  cccccccccccccccccc  cccccccccc Log#5  2 x'  ccccccc Log#1  x2 '
29
5
145
58

Simple Differential Equations

1y
ccc , y+1/ 2, with the help of Mathematica. Separating the
Suppose you want to solve the initial value problem, y' cccccccc
1t2
1
1
c
ccc
c
c
dy
cccccccc
ccc dt. Begin by assigning a name to this differential
variables in the differential equation, you can see that cccc
1y
1t2
equation, after integrating both sides.

In[20]:=

Out[20]=

eq

1
cccccccccccc  y
1y

Log#1  y'

1
 t  c
cccccccccccccc
1  t2

c  ArcTan#t'

Now t and y are replaced with the values 1 and 2, , respectively, using the replacement operator s. and the Solve
command is used to find the value of c.
In[21]:=
Out[21]=

cvalue

Solve#eq s. t 1, y 2, c'

1
c cccc +S  4 Log#3'/
4

So the value of c is cccc14 +S  4ln+3// and this value can be recalled by extracting it from the list with
cvalue##1, 1, 2''.
In[22]:=

cvalue##1, 1, 2''

Out[22]=

1
cccc +S  4 Log#3'/
4

Use the Solve command again to find y after replacing c with the value found in the last output.

Integration

56

In[23]:=

Solve#eq s. c cvalue##1, 1, 2'', y'

sol

y 1  ccc4c
1

Out[23]=

+S4 ArcTan#t'4 Log#3'/

Numerical Integration
2

1
x s2 dx using Mathematica together with the Trapezoidal rule, Midpoint rule or
Suppose you wanted to estimate 0 cccccccc
rcccc
c e
2

2S

Simpson's Rule. First, define the function to be integrated.

In[24]:=

f#x_'

1
x2
cccccccc
cccccc  cccc2cccc ;
r
2S

ba
cccccc and we will let the functions x#i',
In this example, a 0, b 2, and for this example, we will let n 100. Then h cccc
n
y #i', and m#i' represent xi , yi , and mi , respectively. Then using the formulas introduced in Section 5.10 of the text, we
can define the variables used in the numerical integration rules.
In[25]:=

Clear#a, b, n, x, y, h';
n 100; a 0; b 2;
ba
h cccccccccccc ;
n
x#i_' 0  i h;
1
m#i_' cccc +x#i  1'  x#i'/;
2
y#i_' : f#x#i'';
2

1
h
x s2 dx using the Trapezoidal rule, you need to enter and execute cccc
+y#0'  2 n1
To estimate 0 cccccccc
rcccc
c e
i 1 y#i'  y#n'/.
2
2S
2

(The command N #expr, 20' is also used to round the value of expr to 20 digits.)
In[31]:=

Out[31]=

n1
\
h L
]
M
]
M
N$ cccc M

2
y#0'
y#i'  y#n']
M
], 20(
2 N
i 1
^

0.47724626867805128894

Midpoint Rule:
n
In[32]:=

N$h f#m#i'', 20(

i 1

Out[32]=

0.47725166772970608614

Simpson's Rule
In[33]:=

Out[33]=

h
N$ cccc +y#0'  4 Sum#y#i', i, 1, n  1, 2' 
3
2 Sum#y#i', i, 2, n  2, 2'  y#n'/, 20(
0.47724986795579570810

Integration

NIntegrate Command
b

The built-in command, NIntegrate#f #x', x, a, b' is used to numerically integrate a f +x/dx using a built-in
Mathematica algorithm.
In[34]:=
Out[34]=

NIntegrate#f#x', x, 0, 2'

0.47725

Graphing Polar and Parametric Equations

The Add-on package Graphics (in the Graphics directory) contains the command PolarPlot#f #T', T, a, b'
which will plot the polar equation r f +T/ for values of T from a to b.
In[35]:=

 Graphics`Graphics`

In[36]:=

PolarPlot#Sin#11T', T, 0, S'

1

0.5

-1

-0.5

0.5

-0.5

-1
Out[36]=

hGraphicsh

To plot the parametric equation, g +T/;Cos#T', Sin#T'?, execute

ParametricPlot#g #T'Cos#T', Sin#T', T, a, b'.

57

Integration

58

In[37]:=

gr

ParametricPlot#Sin#11 T'Cos#T', Sin#T', T, 0, S, AspectRatio 1'

1

0.5

-1

-0.5

0.5

-0.5

-1
Out[37]=

hGraphicsh

Computing ArcLength
Suppose you wanted to compute the arclength of the curve below for the given parametric equations.
In[38]:=

x#t_' t  Cos#t';
y#t_' t  Sin#2t';
ParametricPlot#x#t', y#t', t, 0, 4'
3
2.5
2
1.5
1
0.5

-1
Out[40]=

hGraphicsh

Integration

59

Using formula (5) in Section 6.4 of the text, you can attempt to find the value of the arclength with Mathematica.
4
In[41]:=



x '#t'2  y '#t'2  t

0
4

Out[41]=




+1  2 Cos#2 t'/2  +1  Sin#t'/2  t

Mathematica is unable to find the exact value of the integral, so you can instead use the NIntegrate command instead (or
you could use the numerical integration methods introduced in Section 5.10).
In[42]:=
Out[42]=

NIntegrate$



x '#t'2  y '#t'2 , t, 0, 4(

9.33017

Here is another example where the arclength of the graph of f +x/

(10) in Section 6.4.
In[43]:=

Out[44]=

sin+x/, from x

0 to x

2S, is computed using formula

Sin#x';

NIntegrate$ 1  f '#x'2 , x, 0, 2 S(

f#x_'

7.6404

Improper Integrals

1
ccccccc dx
Improper integrals are illustrated in this section using a few simple examples. Suppose you want to compute 1 cccccccc
+1x/2
b

1
ccccccc dx and therefore the Limit command can be
using Mathematica. By definition, this integral represents limb 1 cccccccc
+1x/2

used to compute the improper integral.

In[45]:=

Out[46]=

Clear#b, x';
b
1
Limit$ cccccccccccccccccccc  x, b (
2
1 +1  x/
1
cccc
2

Mathematica can also evaluate the improper integral directly if you replace b with .

In[47]:=

Out[47]=

1
cccccccccccccccccccc  x
+1  x/2

1
cccc
2

1
Since 1 cccccccc
rcccc
cccc  x converges, Mathematica is able to evaluate the integral successfully.
x1

In[48]:=

Out[48]=

2
1
cccccccccc
cccccccc
r
  x
1
x1

When you attempt to evaluate the following integral, Mathematica outputs a warning message indicating that the integral
does not converge.

Integration

60

In[49]:=

2
1
 x
cccccccccccccccccccccccc
3s2
1 +x  1/

1
Integrate::idiv : Integral of cccccccccccccccccccccccc
ccc does not converge on 1, 2.
+1  x/3s2
2

Out[49]=

1
ccc  x
cccccccccccccccccccccccc
3s2
1 +1  x/
2

1
cccccccccc  x.
To determine if the integral diverges to or  (or neither), use Mathematica to compute lima1 a cccccccc
+x1/3 2
s

In[50]:=

Clear#a';
2
1
Limit$ cccccccccccccccccccccccc  x, a 1, Direction 1(
3s2
a +x  1/

Out[51]=

Taylor Polynomials and Taylor Series

The Mathematica command Normal#Series#f #x', x, c, n'' will find the Taylor polynomial Tn +x; c/ for the
function f +x/. For example, you can use the following command to find T11 +x;0) for sin+x/.
In[52]:=

Out[53]=

Clear#f';
t#x_' Normal#Series#Sin#x', x, 0, 11''
x3
x5
x7
x9
x11
x  ccccccc  cccccccccc  ccccccccccccc  cccccccccccccccccc  cccccccccccccccccccccccc
6
120
5040
362880
39916800

By graphing the function together the corresponding Taylor polynomial, you can compare the two functions.
In[54]:=

Plot#Sin#x', t#x', x, 3S, 3S, PlotRange 2, 2, PlotStyle

Thickness#.03', RGBColor#0, 1, 0', Thickness#.01', RGBColor#0, 0, 1''
2
1.5
1
0.5
-7.5

-5

-2.5
-0.5

2.5

7.5

-1
-1.5
-2
Out[54]=

hGraphicsh

From the graph above, you can see that the polynomial is a "good" approximation of the sine curve from about 5 to 5. To
see the difference between the two functions, graph sin+x/  T11 +x; 0/ . From the corresponding graph below, it appear
that the difference between sin+x/ and T11 +x; 0/ is less than about 0.175 for 5 x 5.

Integration

In[55]:=

Plot#Abs#Sin#x'  t#x'', x, 5, 5, PlotRange All'

0.175
0.15
0.125
0.1
0.075
0.05
0.025
-4

Out[55]=

-2

hGraphicsh

Here is another example of a Taylor series using c

In[56]:=

S.

Clear#t';
t#x_' Normal#Series#Sin#x', x, S, 11'';
Plot#Sin#x', t#x', x, 3S, 3S, PlotRange 2, 2, PlotStyle
Thickness#.03', RGBColor#0, 1, 0', Thickness#.01', RGBColor#0, 0, 1''
2
1.5
1
0.5
-7.5

-5

-2.5
-0.5

2.5

7.5

-1
-1.5
-2
Out[58]=

hGraphicsh

The command Series#f #x', x, c, n' will display the beginning terms of a Taylor series up to powers of n. The
last term O#x'12 represents all the terms with powers of 12 and higher which are not explicitly shown.
In[59]:=
Out[59]=

Series#Sin#x', x, 0, 11'

x5
x7
x9
x11
x3
x  ccccccc  cccccccccc  ccccccccccccc  cccccccccccccccccc  cccccccccccccccccccccccc  O#x'12
6
120
5040
362880
39916800

61

Integration

62

Sequences and Series

The command Table#a#n', k, 1, n' can be used to generate the first n terms of the sequence a#1', a#2', a#3', ...
sin+1sn/
ccccccccc .
. For example, here are the first 50 terms of the sequence  cccccccc
1sn
n 1

In[60]:=

Out[60]=

1
Sin# ccc
c'
n
Table$ ccccccccccccccccccccc ss N, n, 1, 50(
1
ccc
c
n

0.841471,
0.997398,
0.999259,
0.999656,
0.999802,
0.999871,
0.999914,

0.958851,
0.997944,
0.999349,
0.999685,
0.999815,
0.999878,
0.999918,

0.981584,
0.998334,
0.999423,
0.999711,
0.999827,
0.999885,
0.999921,

0.989616, 0.993347, 0.995377, 0.996602,

0.998623, 0.998843, 0.999014, 0.99915,
0.999486, 0.999538, 0.999583, 0.999622,
0.999733, 0.999753, 0.999771, 0.999787,
0.999837, 0.999847, 0.999856, 0.999864,
0.99989, 0.999896, 0.999901, 0.999906, 0.99991,
0.999925, 0.999928, 0.999931, 0.999933

5
Now, for example, suppose you wish to compute
cccccccc . As indicated in Section 7.4 of the text, the value of
n 1 cccccccc
n+n1/

5
k
5

is
the
limit
of
the
sequence
cccccccc
c
ccc
c
ccc
cccccccc
c
ccc
c
ccc
.
The
output below shows the first 200 terms of this sequence.

n 1 n+n1/
n 1 n+n1/
k 1

In[61]:=

Out[61]=

5
\ ss N, k, 1, 200(
M
Table$ L
M cccccccccccccccccccccc ]
]

n+n
1/
N
^
n 1
2.5, 3.33333, 3.75, 4., 4.16667, 4.28571, 4.375, 4.44444, 4.5, 4.54545,
4.58333, 4.61538, 4.64286, 4.66667, 4.6875, 4.70588, 4.72222, 4.73684,
4.75, 4.7619, 4.77273, 4.78261, 4.79167, 4.8, 4.80769, 4.81481,
4.82143, 4.82759, 4.83333, 4.83871, 4.84375, 4.84848, 4.85294, 4.85714,
4.86111, 4.86486, 4.86842, 4.87179, 4.875, 4.87805, 4.88095, 4.88372,
4.88636, 4.88889, 4.8913, 4.89362, 4.89583, 4.89796, 4.9, 4.90196,
4.90385, 4.90566, 4.90741, 4.90909, 4.91071, 4.91228, 4.91379, 4.91525,
4.91667, 4.91803, 4.91935, 4.92063, 4.92188, 4.92308, 4.92424, 4.92537,
4.92647, 4.92754, 4.92857, 4.92958, 4.93056, 4.93151, 4.93243, 4.93333,
4.93421, 4.93506, 4.9359, 4.93671, 4.9375, 4.93827, 4.93902, 4.93976,
4.94048, 4.94118, 4.94186, 4.94253, 4.94318, 4.94382, 4.94444, 4.94505,
4.94565, 4.94624, 4.94681, 4.94737, 4.94792, 4.94845, 4.94898, 4.94949,
4.95, 4.9505, 4.95098, 4.95146, 4.95192, 4.95238, 4.95283, 4.95327,
4.9537, 4.95413, 4.95455, 4.95495, 4.95536, 4.95575, 4.95614, 4.95652,
4.9569, 4.95726, 4.95763, 4.95798, 4.95833, 4.95868, 4.95902, 4.95935,
4.95968, 4.96, 4.96032, 4.96063, 4.96094, 4.96124, 4.96154, 4.96183,
4.96212, 4.96241, 4.96269, 4.96296, 4.96324, 4.9635, 4.96377, 4.96403,
4.96429, 4.96454, 4.96479, 4.96503, 4.96528, 4.96552, 4.96575, 4.96599,
4.96622, 4.96644, 4.96667, 4.96689, 4.96711, 4.96732, 4.96753, 4.96774,
4.96795, 4.96815, 4.96835, 4.96855, 4.96875, 4.96894, 4.96914,
4.96933, 4.96951, 4.9697, 4.96988, 4.97006, 4.97024, 4.97041, 4.97059,
4.97076, 4.97093, 4.9711, 4.97126, 4.97143, 4.97159, 4.97175, 4.97191,
4.97207, 4.97222, 4.97238, 4.97253, 4.97268, 4.97283, 4.97297,
4.97312, 4.97326, 4.9734, 4.97354, 4.97368, 4.97382, 4.97396, 4.97409,
4.97423, 4.97436, 4.97449, 4.97462, 4.97475, 4.97487, 4.975, 4.97512

Integration

From the output above, the sequence of partial sums appears to converge to 5. To verify this conjecture, the limit is com5
cccccccc 0.
puted to obtain
n 1 , cccccccc
n+n1/
In[62]:=

Out[62]=

k
5
L cccccccccccccccc
\, k ( ss Simplify
cccccc ]
Limit$ M
M
]

n+n
1/
N
^
n 1

Mathematica can also compute the limit directly.

In[63]:=

Out[63]=

5
L cccccccccccccccc
\
cccccc ]
M
M
]

n+n
1/
N
^
n 1

63