Calculus , Johnston & Mathews, MATLAB

Manual

Isom Herron, R.P.I.

Contents
1 Information for Instructors 1
1.1 Getting Started . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Student Accessibility to MATLAB . . . . . . . 2
1.1.2 Working in a MATLAB Environment . . . . . 3

2 Plotting, Limits and the Derivative 5

2.1 Plotting simple functions with MATLAB . . . . . . . 5
2.2 Taking limits with MATLAB . . . . . . . . . . . . . . . 8
2.3 The Derivative with MATLAB . . . . . . . . . . . . . . 9
2.4 Solving Equations by MATLAB . . . . . . . . . . . . . 10
2.5 Plotting Parametric Equations . . . . . . . . . . . . . . 11
2.6 Plotting Polar Equations . . . . . . . . . . . . . . . . . 12
2.7 Approximate solutions of equations . . . . . . . . . . . 12
2.8 Applications of the derivative . . . . . . . . . . . . . . 13

3 Integration 14
3.1 A De1nite Integral . . . . . . . . . . . . . . . . . . . . . 14
3.2 Integration by Parts; An Inde1nite Integral . . . . . 15
3.3 Approximation of Integrals . . . . . . . . . . . . . . . . 16
3.3.1 Trapezoidal Rule . . . . . . . . . . . . . . . . . . 16
3.3.2 Taylor Series . . . . . . . . . . . . . . . . . . . . . 16

1 Information for Instructors

Many of the exercise sets in Calculus by Johnson and Mathews contain
technology exercises, marked with the icon T which are intended to be
 Calculus , J & M 
, MATLAB M  2

solved with the use of a computer algebra system (CAS) such as MAT-
LAB. In addition, each chapter of the text contains at least one student
project. These projects are designed to give students substantial experience
in solving multistep problems, usually with the aid of a graphical calculator
or a CAS. The purpose of this manual is to provide a brief introduction to
the MATLAB computing environment, speci1cally geared to the textbook.
If you are considering having students use MATLAB for some of the
Technology exercises or Student Projects, please spend a few minutes read-
ing this section. It contains general advice and suggestions for the successful
implementation of MATLAB in your course.

1.1 Getting Started

1.1.1 Student Accessibility to MATLAB
First you may need to determine where your students will have access to
MATLAB. For instance, if you plan on having students do their work on
campus in a computer lab, then it will be necessary that MATLAB is readily
available in the lab or labs, where the students will be expected to work.
If possible, the labs should have the latest version of MATLAB (presently
Version 6.1 at the time of writing). Up-to-date information about MAT-
LAB can be found on the web at http://www.mathworks.com. If you or
your school is planning on purchasing MATLAB for use in a computer lab,
the information you may need can also be obtained from the Mathworks
website. An academic site license can be obtained at a considerable dis-
count. Once you purchase and obtain your site license for MATLAB, have
your computer administrator install the program on the lab machines. For
instance, on our campus, MATLAB is available on the computer network
as soon as the students log-in.
If your school chooses not to obtain a site license, you may need to
make arrangements to have your students purchase individual copies of
MATLAB. The Student Edition of MATLAB is recommended. It is de-
scribed at the URL
http://www.mathworks.com/products/student version/. You may want to
check to see if the campus bookstore or computer store will stock the soft-
ware for you.
You might want to spend some time working with MATLAB in the same
environment as your students. If they will be working in an on-campus lab,
 Calculus , J & M 
, MATLAB M  3

spend some time in the lab. Learn any special access procedures and how
to launch MATLAB. Check to see if you can access materials in a course
archive or on the WWW, save a workspace 1le, exit MATLAB, transfer a
1le via e-mail or onto a diskette, and log out of the computer.

1.1.2 Working in a MATLAB Environment

For a vast range of applications, academic and non-academic, MATLAB
has become the technical computing environment of choice. To fully survey
all of its possible applications would take us far beyond the scope of the
textbook and that of this guidebook. However, MATLAB was designed
to be user-friendly for mathematical calculations involving matrices and
vectors. Its use for solving calculus problems has great power and appeal
as well.
The code described in this manual was written speci1cally to solve the
problems in the text. It was tested in an X window system, both locally
and remotely. It was also run in the student edition of MATLAB on a
Macintosh machine.
Once MATLAB is opened the window which appears is the command
window. The command line is the input line next to the MATLAB prompt.
Navigating the command line will depend on the system one uses, but most
will permit editing lines since speci1c MATLAB syntax must be observed.
The MATLAB CAS commands are contained in the Symbolic Math
Toolbox. MATLAB will generally produce a numerical answer to a cal-
culation unless a symbolic answer is desired. The command help symbolic
will give a list of the symbolic functions available. They range from Cal-
culus, Linear Algebra, Solutions of Differential and Functional Equations,
Integral Transforms, Conversions and Plots. Use the command demo to
explore the wide range of applications available in MATLAB.
MATLAB is truly a language, and a sequence of commands used over
and over again can be composed in a program called an M
file. There
are two types of M -f iles, scripts and functions. The use of an M -f ile
may become necessary when MATLAB, on a rare occasion, does not have
a built-in command or function to solve a problem in the text.
The commands used in this manual are listed below:
 Calculus , J & M 
, MATLAB M  4

Command Function
syms Short cut for constructing symbolic objects
ezplot Easy-to-use function plotter
limit Limit of a symbolic expression
solve Symbolic solution of functional equations
double Convert symbolic expression to numerical
plot Linear plot
linspace Linearly spaced vector
subs Symbolic substitution
diff Differentiate
int Integrate
pretty Pretty print a symbolic expression
taylor Taylor series
polar Plot in polar coordinates
trapz Trapezoidal integration
 Calculus , J & M 
, MATLAB M  5

2 Plotting, Limits and the Derivative

2.1 Plotting simple functions with MATLAB

One of the most useful bene1ts of a program such as MATLAB is in plotting
functions. There are several ways in which this may be carried out. The
most common command for two-dimensional graphs is plot. Data arrays
are formed and plotted on the horizontal and vertical axes, and the points
are connected by line segments. If there are not enough line segments,
the curve may appear choppy. The command to segment the interval is
linspace, which is actually an internal function, since it is made up of more
than one command. You may adjust the number of data points as needed.
However there is a default number of 100, which is often adequate. To
perform a plot, the process is illustrated on a T exercise from Section 1.2,
p.19, #39.

>> syms x y y1 y2
>> x = linspace(0, 2  pi);
>> y = sin(x);
>> y1 = sin(y);
>> y2 = sin(y1);
>> plot(x, y)
 Calculus , J & M 
, MATLAB M  6

>> plot(x, y, x, y1)

>>

>> plot(x, y, x, y1, x, y2 )

>>
 Calculus , J & M 
, MATLAB M  7

Alternatively, the command ezplot may be used to perform a plot. This

is particularly recommended when only one function over one interval is
needed. This is illustrated on a T exercise from Section 1.4, p.35, #13.

>> syms x y
>> y = 2ˆx;
>> ezplot(y, [
1 5])
>>

This shows the graph of y = 2x for
1 < x < 5.

 Calculus , J & M 
, MATLAB M  8

>> ezplot(y, [
0.1, 0.1])

This shows zooming in near the point (0, 1).

2.2 Taking limits with MATLAB

Just about all of the important processes of calculus may be performed
using MATLAB. The 1rst one to be shown is that of taking limits. This is
illustrated on a T exercise in section 1.5, p.47, #13 of the text. If no limit
point is given, it is automatically taken to be 0.

>> syms h
>> limit(sin(h)/h)
ans =
1

Another possibility is that the limit does not exist. MATLAB

identi1es this difficulty, which is also on p.47 in the text.

>> syms x
>> limit(tan(x), pi/2)
 Calculus , J & M 
, MATLAB M  9

ans =
NaN
>>help NaN
NaN Not-a-Number.
NaN is the IEEE arithmetic representation for Not-a-Number.
A NaN is obtained as a result of mathematically undefined
operations like 0.0/0.0 and inf-inf.

A more complicated example of a limit that does exist is the following

T exercise from section 1.6, p.63, #17.

>> limit((tan(x)
1)/(x
pi/4), pi/4)

ans =
2

It may also be necessary to take a one-sided limit. MATLAB will also

evaluate such a limit as is illustrated on an example from Section 1.6, p.63.
The 1rst place after the function is the variable v which is to approach the
limit c from the left, given in single quotes.

>> syms v c
>> limit(sqrt(1
(v/c)ˆ2), v, c, left )
ans =
0

2.3 The Derivative with MATLAB

The de1nition of the derivative is based on taking a limit. Hence it is very
natural to apply the techniques just used. The derivation of the derivative
of cos(x) follows easily from its de1nition in Example 1 on p.123:
d cos(x + h)
cos(x)
cos(x) =lim .
dx h→0 h
The MATLAB derivation is as follows.

>> limit((cos(x + h)
cos(x))/h, h, 0)

ans =
 Calculus , J & M 
, MATLAB M  10

-sin(x)

The derivatives of other functions such as bx may also be derived. This

is motivated on p.134 in the text.

>>syms h b
>>limit((b^h-1)/h,h,0)

ans =

log(b)
>> % The log is the natural logarithm ln.

2.4 Solving Equations by MATLAB

The command solve(Ef (x) = 0E) in MATLAB solves the equation f (x) = 0
for the desired values of x. As the need arises this command will be greatly
used. A 1rst illustration is from section 2.7, a T exercise on p.152, #43.
The command gives nine roots of the equation 2x = x8 , in terms of the built-
in function lambertw, which at this stage is not too useful to us. MATLAB,
the genie that it is, desires to please us by obtaining as exact an answer as
possible.

>>[x] =solve( 2^x=x^8 )

x=

[
8  lambertw(
1/8  log(2))/log(2)]

[
8  lambertw(
1,
1/8  log(2))/log(2)]
[
8  lambertw(
1/8  log(2)  (1/2  2ˆ(1/2) + 1/2  i  2ˆ(1/2)))/log(2)]
[
8  lambertw(
1/8  i  log(2))/log(2)]
[
8  lambertw(
1/8  log(2)  (
1/2  2ˆ(1/2) + 1/2  i  2ˆ(1/2)))/log(2)]
[
8  lambertw(1/8  log(2))/log(2)]
[
8  lambertw(
1/8  log(2)  (
1/2  2ˆ(1/2)
1/2  i  2ˆ(1/2)))/log(2)]
[
8  lambertw(1/8  i  log(2))/log(2)]
[
8  lambertw(
1/8  log(2)  (1/2  2ˆ(1/2)
1/2  i  2ˆ(1/2)))/log(2)]
However, MATLAB, being the practical tool it is, has another command,
double which converts a symbolic expression to a numerical one.
 Calculus , J & M 
, MATLAB M  11

>> double(x)
ans =
1.1000
43.5593
0.6972 + 0.8014i

0.0850 + 0.9890i

0.7007 + 0.6283i

0.9231

0.7007
0.6283i

0.0850
0.9890i
0.6972
0.8014i
Only the positive real values of x are of interest to us. Thus for 0 < x <
1.1, x8 < 2x , while for 1.1 < x < 43.56, 2x < x8 , but for x > 43.56, 2x > x8 .

2.5 Plotting Parametric Equations

The need may arise to plot a curve given in parametric form. The plot
command is capable of handling this, as illustrated on a T exercise from
section 3.4, p.232, #79.
>>t=0:0.1:7;
>> x=sin(t);
>>y=sin(2*t);
>>plot(x,y)
 Calculus , J & M 
, MATLAB M  12

2.6 Plotting Polar Equations

Plots in polar coordinates can be done using the command polar(t, r), where
t is the angle in radians and r is the radius vector. This is illustrated on a
T exercise from section 6.3, p.471, #39.
>> % Graph the curve described by r=ln(theta), 0< theta< 2*pi
>>t=linspace(0,2*pi);
>>r=log(t);
Warning: Log of zero.
>>polar(t,r)
>>title( r=ln(theta) )
>>
>>

2.7 Approximate solutions of equations

NewtonEs method is fast and accurate in 1nding the solution of the equa-
tion f (x) = 0 when a reasonably good initial guess is known. The solve
command will give the desired answers. This is illustrated on a T exercise
in Section 4.2, on p.275, #35.

>>solve( x^4-5*x^3+21*x^2+13*x+49=0 );
 Calculus , J & M 
, MATLAB M  13

>>double(ans)
ans =
2.9956- 4.0003i
2.9956+ 4.0003i
-0.4956+ 1.3101i
-0.4956- 1.3101i

2.8 Applications of the derivative

The command diff performs the symbolic differentiation of a function. This
technique is illustrated on an interesting but complicated T exercise from
section 4.7, p.319, #23.
>>syms A r h
>>A=pi^r^2/2-r^2*asin(1-h/r)-(r-h)*sqrt(h*(2*r-h))

A =

1/2*(pi^r)^2-r^2*asin(1-h/r)-(r-h)*(h*(2*r-h))^(1/2)

>>Ah=diff(A,h)

Ah=
r/(1-(1-h/r)^2)^(1/2)+(h*(2*r-h))^(1/2)
-1/2*(r-h)/(h*(2*r-h))^(1/2)*(2*r-2*h)

>>Ah1=subs(Ah,r,1.5)

Ah1=
3/2/(1-(1-2/3*h)^2)^(1/2)+(h*(3-h))^(1/2)
-1/2*(3/2-h)/(h*(3-h))^(1/2)*(3-2*h)

>>Ah2=subs(Ah1,h,2)

Ah2 =

3/16*8^(1/2)*9^(1/2)+7/8*2^(1/2)

>>hrate=0.5/(6*Ah2)
 Calculus , J & M 
, MATLAB M  14

hrate =

1/2/(9/8*8^(1/2)*9^(1/2)+21/4*2^(1/2))

>>double(hrate)
ans =
0.0295
>>

3 Integration

3.1 A De1nite Integral

The command int may be used to obtain both inde1nite and de1nite inte-
grals. Whenever possible, MATLAB attempts to provide an exact answer.
This is illustrated on a T example from section 5.6, p.393, #35. This
example is somewhat involved and the code is developed next.
>>% First, find the area of the region removed. It is determined
by the points where the circle and the parabola meet.
>>syms x
>>solve( 4-sqrt(16-x^2)=-0.4*x^2+3.2 )

ans =

[ -2.4356475028411506249717241735830]
[ 2.4356475028411506249717241735830]

>>% The area of the removed section is then obtained as a definite

integral between those limits.
>>A=int(-0.4*x^2+3.2-(4-sqrt(16-x^2)),x,-2.4356475,2.4356475)

A =

-1037004934973087065929065648105004366319763483/133804471
191183738849214309635890169035882496+4113442940027205/31691
 Calculus , J & M 
, MATLAB M  15

2650057057350374175801344*354506281325390153376730559031
^(1/2)+16*asin(1371147646675735/2251799813685248)
>>double(A)
ans =
10.4530
>> % The remaining area, and the mass, are given in the following
calculations.
>>Ar=16*pi-10.4530
Ar =
39.8125
>>M=Ar*0.5*8920*1e-6
M =
0.17756
>>

3.2 Integration by Parts; An Inde1nite Integral

MATLAB will quickly evaluate inde1nite or de1nite integrals which, if done
by hand, would require lengthy algebraic or trigonometric calculations. For
instance, here is the MATLAB calculation for Example 5 in section 5.7:
>>syms t
>>A=int(exp(t)*(65+10*cos(pi*t/12)-5*cos(pi*t/6)-15*sin(pi*t/12)),t)

A =

65*exp(t)+10/(1+1/144*pi^2)*exp(t)*cos(1/12*pi*t)
+5/6*pi/(1+1/144*pi^2)*exp(t)*sin(1/12*pi*t)
-5/(1+1/36*pi^2)*exp(t)*cos(1/6*pi*t)
-5/6*pi/(1+1/36*pi^2)*exp(t)*sin(1/6*pi*t)
+5/4*pi/(1+1/144*pi^2)*exp(t)*cos(1/12*pi*t)
-15/(1+1/144*pi^2)*exp(t)*sin(1/12*pi*t)
>>
>>% The answer is easier to read if we pretty it up.
>>pretty(A)
exp(t) cos(1/12 pi t) pi exp(t) sin(1/12 pi t)
--------------------- -----------------------
65 exp(t) + 10 + 5/6
2 2
1 + 1/144 pi 1 + 1/144 pi
 Calculus , J & M 
, MATLAB M  16

exp(t) cos(1/6 pi t) pi exp(t) sin(1/6 pi t)

-------------------- -----------------------
- 5 2 - 5/6 2
1 + 1/36 pi 1 + 1/36 pi
pi exp(t) cos(1/12 pi t) exp(t) sin(1/12 pi t)
------------------------ ---------------------
+ 5/4 2 - 15 2
1 + 1/144 pi 1 + 1/144 pi

3.3 Approximation of Integrals

3.3.1 Trapezoidal Rule
There are three built-in MATLAB functions for numerically evaluating an
integral over a 1nite range. They are trapz , quad and quad8. The trapezoid
rule is employed in the 1rst of those commands. It is illustrated in solving
a T exercise from section 6.4 on p.479, #27. The purpose is to evaluate

/2
0 (a2
b2 ) sin 2 x + b2 dx. Since the default number in linspace is 100,
the effect is to compute T100 .
>>x=linspace(0,pi/2);
>>a=17.9435;
>>b=4.5511;
>>
>>f=sqrt((a^2-b^2)*(sin(x).^2.)+b^2);
>>trapz(x,f)
ans =
19.2710

3.3.2 Taylor Series

MATLAB can be used to 1nd the Taylor series of a function. The com-
mand taylor(f ) gives (by default) the 1fth order Maclaurin polynomial
approximation to f , that is, the Taylor series about x = 0. Here we em-
ploy taylor (f, n) which is the (n
1)-st order Maclaurin polynomial. Several
other parameters can be speci1ed and a complete description may be found
under help taylor. For very smooth functions, the Taylor series method will
 Calculus , J & M 
, MATLAB M  17

give an accurate approximation to a de1nite integral, using few terms when

the interval of integration is short. The method is developed in an example
at the end of section 7.7 on p.621 in the text. In what follows the method
is illustrated on a T exercise, p.624, #21.
>>syms f x
>>f=taylor(sin(x^2),15)

f =

x^2-1/6*x^6+1/120*x^10-1/5040*x^14

>>I=int(f,x,0,pi/2)

I =

1/24*pi^3-1/5376*pi^7+1/2703360*pi^11-1/2477260800*pi^15

>>double(I)
ans =
0.8274