MATLAB Differential and Integral Calculus

www.pdfgrip.
com
For your convenience Apress has placed some of the front

matter material after the index. Please use the Bookmarks
and Contents at a Glance links to access them.
www.pdfgrip.com
Contents at a Glance
About the Author��ix
■■Chapter 1: Introduction and the MATLAB Environment��1

■■Chapter 2: Limits and Continuity. One and Several Variables��17
■■Chapter 3: Numerical Series and Power Series��45
■■Chapter 4: Derivatives and Applications. One and Several Variables��69
■■Chapter 5: Vector Differential Calculus and Theorems in Several Variables��109
■■Chapter 6: Integration and Applications��129
■■Chapter 7: Integration in Several Variables and Applications��163
■■Chapter 8: Differential Equations��181
iii
www.pdfgrip.com
Chapter 1
Introduction and the MATLAB

Environment
1.1 Numerical Computation with MATLAB

You can use MATLAB as a powerful numerical computer. While most calculators handle numbers only to a preset
degree of precision, MATLAB performs exact calculations to any desired degree of precision. In addition, unlike
calculators, we can perform operations not only with individual numbers, but also with objects such as arrays.
Most of the topics of classical numerical analysis are treated by this software. It supports matrix calculus, statistics,
interpolation, least squares fitting, numerical integration, minimization of functions, linear programming, numerical
and algebraic solutions of differential equations and a long list of further methods that we’ll meet as this book progresses.
Here are some examples of numerical calculations with MATLAB. (To obtain the results simply press Enter once
the desired command has been entered after the prompt “>>”.)
1. We calculate 4 + 3 to obtain the result 7. To do this, just type 4 + 3, and then Enter.

>> 4 + 3

ans =

7

2. We find the value of 3 to the power of 100, without having previously set the precision.
To do this we simply enter 3 ^ 100.

>> 3 ^ 100

ans =

5. 1538e + 047

3. We can use the command “format long e” to obtain results to 15 digits (floating-point).

>> format long e

>> 3^100

ans =

5.153775207320115e + 047

1
www.pdfgrip.com
Chapter 1 ■ Introduction and the MATLAB Environment
4. We can also work with complex numbers. We find the result of the operation raising
(2 + 3i) to the power 10 by typing the expression (2 + 3i) ^ 10.

>> (2 + 3i) ^ 10

ans =

-1 415249999999998e + 005 - 1. 456680000000000e + 005i

5. The previous result is also available in short format, using the “format short” command.

>> format short
>> (2 + 3i)^10

ans =

-1.4152e + 005- 1.4567e + 005i

6. We can calculate the value of the Bessel function J0 at 11.5. To do this we type
besselj(0,11.5).

>> besselj(0,11.5)

ans =

-0.0677

7. We can also perform numerical integration. To calculate the integral of sin(sin(x)) between
0 and p we type int(‘sin((sin(x))’, 0, pi).

>> int ('sin(sin(x))', 0, pi)

ans =

1235191162052677/2251799813685248 * pi

These ideas will be treated more thoroughly later in the book.
1.2 Symbolic Computation with MATLAB

MATLAB perfectly handles symbolic mathematical computations, manipulating and performing operations on
formulae and algebraic expressions with ease. You can expand, factor and simplify polynomials and rational and
trigonometric expressions, find algebraic solutions of polynomial equations and systems of equations, evaluate
derivatives and integrals symbolically, find solutions of differential equations, manipulate powers, and investigate
limits and many other features of algebraic series.
To perform these tasks, MATLAB first requires all the variables (or algebraic expressions) to be written between
single quotes. When MATLAB receives a variable or expression in quotes, it is interpreted as symbolic.
Here are some examples of symbolic computations with MATLAB.
2
www.pdfgrip.com
1. We can expand the following algebraic expression: ((x + 1)(x + 2) - (x + 2) ^ 2)^3. This
is done by typing: expand(‘((x + 1)(x + 2) - (x + 2) ^ 2) ^ 3’). The result will be another
algebraic expression:

>> syms x; expand(((x + 1) *(x + 2)-(x + 2) ^ 2) ^ 3)

ans =

-x ^ 3-6 * x ^ 2-12 * x-8

2. We can factor the result of the calculation in the above example by typing:
factor(‘((x + 1) *(x + 2) - (x + 2) ^ 2) ^ 3’)

>> syms x; factor(((x + 1)*(x + 2)-(x + 2)^2)^3)

ans =

-(x+2)^3

3. We can find the indefinite integral of the function (x ^ 2) sin(x) ^ 2 by typing:
int(‘x ^ 2 * sin(x) ^ 2’, ‘x’)

>> int('x^2*sin(x)^2', 'x')

ans =

x ^ 2 *(-1/2 * cos(x) * sin(x) + 1/2 * x)-1/2 * x * cos(x) ^ 2 + 1/4 * cos(x) * sin(x) +
1/4 * 1/x-3 * x ^ 3

4. We can simplify the previous result:

>> syms x; simplify(int(x^2*sin(x)^2, x))

ans =

sin(2*x)/8 -(x*cos(2*x))/4 -(x^2*sin(2*x))/4 + x^3/6

5. We can present the previous result using a more elegant mathematical notation:

>> syms x; pretty(simplify(int(x^2*sin(x)^2, x)))

ans =
2 3
sin(2 x) x cos(2 x) x sin(2 x) x
-------- - ---------- - ----------- + --
8 4 4 6

3
www.pdfgrip.com
6. We can find the series expansion up to order 12 of the function x ^ 2 * sin (x) ^ 2,
presenting the result in elegant form:

>> pretty(taylor('x^2*sin(x)^2',12))

4 6 8 10 12
x - 1/3 x + 2/45 x - 1/315 x + o (x)

7. We can solve the equation 3ax - 7 x ^ 2 + x ^ 3 = 0 (where a is a parameter):

>> solve('3*a*x-7*x^2 + x^3 = 0', 'x')

ans =

[ 0]
[7/2 + 1/2 *(49-12*a) ^(1/2)]
[7/2-1/2 *(49-12*a) ^(1/2)]

8. We can find the five solutions of the equation x ^ 5 + 2 x + 1 = 0:

>> solve('x^5+2*x+1','x')

ans =

RootOf(_Z^5+2*_Z+1)

As the result does not explicitly give five solutions, we apply the “allvalues” command:

>> allvalues(solve('x^5+2*x+1','x'))

ans =

[-.7018735688558619-. 8796971979298240 * i]
[-. 7018735688558619 +. 8796971979298240 * i]
[-. 4863890359345430]
[.9450680868231334-. 8545175144390459 * i]
[. 9450680868231334 +. 8545175144390459 * i]

On the other hand, MATLAB can use the Maple program libraries to work with symbolic math, and can thus
extend its field of action. In this way, MATLAB can be used to work on such topics as differential forms, Euclidean
geometry, projective geometry, statistics, etc.
At the same time, Maple can also benefit from MATLAB’s powers of numerical calculation, which might be used,
for example, in combination with the Maple libraries (combinatorics, optimization, number theory, etc.)
1.3 MATLAB and Maple

Provided the “Extended Symbolic Math Toolbox” is installed then MATLAB can extend its symbolic calculation
abilities by making use of the Maple libraries. To use a Maple command from MATLAB, use the command ‘maple’
followed by the corresponding Maple syntax.
4
www.pdfgrip.com
To use a Maple command from Matlab, the syntax is as follows:

maple ('Maple_command_syntax')

or alternatively:

maple 'Maple_command_syntax'

To use a Maple command with N arguments from Matlab, the syntax is as follows:

maple('Maple_command_syntax', argument1,
argument2,..., argumentN)

Here are some examples.
1. We can calculate the limit of the function (x ^ 3-1) / (x-1) as x tends to 1:

>> maple('limit((x^3-1)/(x-1),x=1)')

ans =

3

We could also have used the following syntax:

>> maple limit('(x^3-1)/(x-1),x=1)';

ans =

3

2. We can calculate the greatest common divisor of 10,000 and 5,000:

>> maple('gcd', 10000, 5000)

ans =

5000
1.4 Graphics with MATLAB

MATLAB can generate two- and three-dimensional graphs, as well as contour and density plots. You can graphically
represent data lists, controlling colors, shading and other graphics features. Animated graphics are also supported.
Graphics produced by MATLAB are portable to other programs.
Some examples of MATLAB graphics are given below.
1. We can represent the function xsin(1/x) for x ranging between -p/4 and p/4, taking 300
equidistant points in the interval. See Figure 1-1.

>> x = linspace(-pi/4,pi/4,300);
>> y=x.*sin(1./x);
>> plot(x,y)

5
www.pdfgrip.com
Figure 1-1.
2. We can give the above graph a title and label the axes, and we can add a grid. See Figure 1-2.

>> x = linspace(-pi/4,pi/4,300);
>> y=x.*sin(1./x);
>> plot(x,y);
>> grid;
>> xlabel('Independent variable X');
>> ylabel('Dependent variable Y');
>> title('The function y=xsin(1/x)')

6
www.pdfgrip.com
Figure 1-2.
3. We can generate a graph of the surface defined by the function z = sin(sqrt(x^2+y^2)) /

sqrt(x^2+y^2), where x and y vary over the interval (–7.5, 7.5), taking equally spaced points
0.5 apart. See Figure 1-3.

>> x =-7.5:. 5:7.5;
>> y = x;
>> [X, Y] = meshgrid(x,y);
>> Z=sin(sqrt(X.^2+Y.^2))./sqrt(X.^2+Y.^2);
>> surf(X, Y, Z)

7
www.pdfgrip.com
Figure 1-3.
These 3D graphics allow you to get a clear picture of figures in space, and are very helpful in visually identifying
intersections between different bodies, and in generating all kinds of space curves, surfaces and volumes of revolution.
4. We can generate the three dimensional graph corresponding to the helix with parametric
coordinates: x = sin(t), y = cos(t), z = t. See Figure 1-4.

>> t=0:pi/50:10*pi;
>> plot3(sin(t),cos(t),t)
Figure 1-4.
8
www.pdfgrip.com
5. We can represent a planar curve given by its polar coordinates r = cos(2t) * sin(2t) for
t varying in the range between 0 and p by equally spaced points 0.01 apart. See Figure 1-5.

>> t = 0:. 1:2 * pi;
>> r = sin(2*t). * cos(2*t);
>> polar(t,r)
Figure 1-5.
6. We can make a graph of a symbolic function using the command “ezplot”. See Figure 1-6.

>> y ='x ^ 3 /(x^2-1)';
>> ezplot(y,[-5,5])
Figure 1-6.
We will see more on these concepts later.
9
www.pdfgrip.com
1.5 Environment and General Notation

As for any program, the best way to learn MATLAB is to use it. By practicing on examples you become familiar with
the syntax and notation peculiar to MATLAB. Each example we give consists of the header with the user input prompt
“>>” followed by the MATLAB response on the next line. See Figure 1-7.
Figure 1-7.
At other times, depending on the type of entry (user input) given to MATLAB, the response is returned using the
expression “ans =”. See Figure 1-8.
10
www.pdfgrip.com
Figure 1-8.
It is important to pay attention to the use of uppercase versus lowercase letters, parentheses versus square
brackets, spaces and punctuation (particularly commas and semicolons).
1.6 Help with Commands

You can get help using MATLAB’s drop down menus, but, in addition, support can also be obtained via commands
(instructions or functions), implemented as MATLAB objects.
You can use the help command to get immediate access to diverse information.

>> help

HELP topics:

matlab\general - General purpose commands.
matlab\ops - Operators and special characters.
matlab\lang - Programming language constructs.
matlab\elmat - Elementary matrices and matrix manipulation.
matlab\elfun - Elementary math functions.
matlab\specfun - Specialized math functions.
matlab\matfun - Matrix functions - numerical linear algebra.
11
www.pdfgrip.com
matlab\datafun - Data analysis and Fourier transforms.

matlab\polyfun - Interpolation and polynomials.
matlab\funfun - Function functions and ODE solvers.
matlab\sparfun - Sparse matrices.
matlab\graph2d - Two dimensional graphs.
matlab\graph3d - Three dimensional graphs.
matlab\specgraph - Specialized graphs.
matlab\graphics - Handle Graphics.
matlab\uitools - Graphical user interface tools.
matlab\strfun - Character strings.
matlab\iofun - File input/output.
matlab\timefun - Time and dates.
matlab\datatypes - Data types and structures.
matlab\winfun - Windows Operating System Interface Files(DDE/ActiveX)
matlab\demos - Examples and demonstrations.
toolbox\symbolic - Symbolic Math Toolbox.
toolbox\tour - MATLAB Tour
toolbox\local - Preferences.

For more help on directory/topic, type "help topic".

As we can see, the help command displays a list of program directories and their contents. Help on any given
topic topic can be displayed using the command help topic. For example:

>> help inv

INV Matrix inverse.
INV(X) is the inverse of the square matrix X.
A warning message is printed if X is badly scaled or nearly singular.

See also SLASH, PINV, COND, CONDEST, NNLS, LSCOV.

Overloaded methods
help sym/inv.m

>> help matlab\elfun

Elementary math functions.

Trigonometric.

sin - Sine.
sinh - Hyperbolic sine.
asin - Inverse sine.
asinh - Inverse hyperbolic sine.
cos - Cosine.
cosh - Hyperbolic cosine.
acos - Inverse cosine.
acosh - Inverse hyperbolic cosine.
tan - Tangent.
tanh - Hyperbolic tangent.
12
www.pdfgrip.com
atan - Inverse tangent.

atan2 - Four quadrant inverse tangent.
atanh - Inverse hyperbolic tangent.
sec - Secant.
sech - Hyperbolic secant.
asec - Inverse secant.
asech - Inverse hyperbolic secant.
csc - Cosecant.
csch - Hyperbolic cosecant.
acsc - Inverse cosecant.
acsch - Inverse hyperbolic cosecant.
cot - Cotangent.
coth - Hyperbolic cotangent.
acot - Inverse cotangent.
acoth - Inverse hyperbolic cotangent.

Exponential.

exp - Exponential.
log - Natural logarithm.
log10 - Common(base 10) logarithm.
log2 - Base 2 logarithm and dissect floating point number.
pow2 - Base 2 power and scale floating point number.
sqrt - Square root.
nextpow2 - Next higher power of 2.

Complex.
abs - Absolute value.

angle - Phase angle.
conj - Complex conjugate.
imag - Complex imaginary part.
real - Complex real part.
unwrap - Unwrap phase angle.
isreal - True for real array.
cplxpair - Sort numbers into complex conjugate pairs.

Rounding and remainder.

fix - Round towards zero.
floor - Round towards minus infinity.
ceil - Round towards plus infinity.
round - Round towards nearest integer.
mod - Modulus(signed remainder after division).
rem - Remainder after division.
sign - Signum.

There is a command for help on a certain sequence of characters (lookfor string) which allows you to find all
those functions or commands that contain or refer to the given string string. This command is very useful when there
is no direct support for the specified string, or if you want to view the help for all commands related to the given
sequence. For example, if we seek help for all commands that contain the sequence complex, we can use the lookfor
complex command to see which commands MATLAB provides.

13
www.pdfgrip.com
>> lookfor complex

ctranspose.m: %' Complex conjugate transpose.
CONJ Complex conjugate.
CPLXPAIR Sort numbers into complex conjugate pairs.
IMAG Complex imaginary part.
REAL Complex real part.
CDF2RDF Complex diagonal form to real block diagonal form.
RSF2CSF Real block diagonal form to complex diagonal form.
B5ODE Stiff problem, linear with complex eigenvalues(B5 of EHL).
CPLXDEMO Maps of functions of a complex variable.
CPLXGRID Polar coordinate complex grid.
CPLXMAP Plot a function of a complex variable.
GRAFCPLX Demonstrates complex function plots in MATLAB.
ctranspose.m: %TRANSPOSE Symbolic matrix complex conjugate transpose.
SMOKE Complex matrix with a "smoke ring" pseudospectrum.
1.7 Commands to Exit and Escape to the MS-DOS Environment

There are three ways you can escape from the MATLAB Command Window to the MS-DOS operating system
environment in order to run temporary assignments. Entering the command ! dos_command in the Command
Window allows you to run the specified DOS command in the MATLAB environment. For example:

! dir

The volume of drive D has no label
The volume serial number £ is 145 c-12F2
Directory of D:\MATLAB52\bin

. <DIR> 13/03/98 0:16 .
.. <DIR> 13/03/98 0:16 ..
BCCOPTS BAT 1.872 19/01/98 14:14 bccopts.bat
CLBS110 DLL 219.136 21/08/97 22:24 clbs110.dll
CMEX BAT 2.274 13/03/98 0:28 cmex.bat
COMPTOOL BAT 34.992 19/01/98 14:14 comptool.bat
DF50OPTS BAT 1.973 19/01/98 14:14 df50opts.bat
FENG DLL 25.088 18/12/97 16:34 feng.dll
FMAT DLL 16.896 18/12/97 16:34 fmat.dll
FMEX BAT 2.274 13/03/98 0:28 fmex.bat
LICENSE DAT 470 13/03/98 0:27 license.dat
W32SSI DLL 66.560 02/05/97 8:34 w32ssi.dll
10 file(s) 11.348.865 bytes
directory(s) 159.383.552 bytes free

The command ! dos_command & is used to execute the specified DOS command in background mode. The
command is executed by opening a DOS environment window on the MATLAB Command Window, as shown in
Figure 1-9. To return to the MATLAB environment simply right-click anywhere in the Command Window (the DOS
environment window will close automatically). You can return to the DOS window at any time to run any operating
system command by clicking the icon labeled MS-DOS symbol at the bottom of the screen.
14
www.pdfgrip.com
Figure 1-9.
The command >>dos_command is used to execute the DOS command in the MATLAB screen. Using the three
previous commands, not only DOS commands, but also all kinds of executable files or batch tasks can be executed.
The command >>dos dos_command is also used to execute the specified DOS command in automatic mode in
the MATLAB Command Window.
To exit MATLAB, simply type quit in the Command Window, and then press Enter.
1.8 MATLAB and Programming

By properly combining all the objects defined in MATLAB, according to the rules of syntax of the program, you can
build useful mathematical programming code. Programs usually consist of a series of instructions in which values are
calculated, are assigned names and are reused in further calculations.
As in programming languages like C or FORTRAN, in MATLAB you can write programs with loops, control flow and
conditionals. MATLAB can write procedural programs, i.e., it can define a sequence of standard steps to run. As in C or
Pascal, a Do, For, or While loop can be used for repetitive calculations. The language of MATLAB also includes conditional
constructs such as If–Then–Else. MATLAB also supports different logical operators, such as AND, OR, NOT and XOR.
MATLAB supports procedural programming (with iterative processes, recursive functions, loops, etc.), functional
programming and object-oriented programming. Here are two simple examples of programs. The first generates the
Hilbert matrix of order n, and the second calculates all the Fibonacci numbers less than 1000.

% Generates the Hilbert matrix of order n
t = '1/(i+j-1)';
for i = 1:n
for j = 1:n
a(i,j) = eval(t);
end
end

% Calculates the Fibonacci numbers
f = [1 1]; i = 1;
while f(i) + f(i-1) < 1000
f(i+2) = f(i) + f(i+1);
i = i+1
end
15
www.pdfgrip.com
Chapter 2
Limits and Continuity.

One and Several Variables
MATLAB provides commands that allow you to calculate virtually all types of limits. The same functions are used to
calculate limits of sequences and limits of functions. The commands for the analysis of one and several variables are
similar. In this chapter we will present numerous exercises which illustrate MATLAB’s capabilities in this field.
The syntax of the commands concerning limits are presented below:
maple (‘limit(sequence, n=infinity)’) or limit(sequence, n, inf) or limit(sequence, inf)
calculates the limit as n tends to infinity of the sequence defined by its general term.
maple (‘limit(function, x=a)’) or limit(function, x, a) or limit(function, a) calculates
the limit of the function of the variable x, indicated by its analytical expression, as the
variable x tends towards the value a.
maple (‘limit(function, x=a, right)’) or limit(function, x, a, ‘right’) calculates the limit
of the function of the variable x, indicated by its analytical expression, as the variable x
tends to the value a from the right.
maple (‘limit(function, x=a, left)’) or limit(function, x, a, ‘left’) calculates the limit of
the function of the variable x, indicated by its analytical expression, as the variable x
tends to the value a from the left.
maple (‘limit(expr, var=a, complex)’) is the complex limit of expr as the variable var
tends to the value a.
maple (‘limit(expr, {v1=a1, …, vn=an})’) is the n-dimensional limit of expr as v1 tends
to a1, v2 tends to a2,…, vn tends to an.
maple (‘Limit(expr, var=a)’) or maple (‘Limit(expr, {v1=a1, …, vn=an})’) is the inert
limit of the expression expr for the specified values of the variable or variables.
17
www.pdfgrip.com
Chapter 2 ■ Limits and Continuity. One and Several Variables
2.1 Limits of Sequences

We present some exercises on the calculation of limits of sequences.
Exercise 2-1
Calculate the following limits:
æ -3 + 2n ö
4
1 + 7n 2 + 3n 3 é 1 + n ö4 1 + n ù 1+ n
lim ç ÷ , lim , lim êæç ÷ 5 ú
, lim n 2
n®¥ -7 + 3n n®¥ 5 - 8n + 4n 3
è ø n®¥
ëêè 2 ø n úû n®¥ n
In the first two limits we face the typical uncertainty given by the quotient ¥ ¥ :

>> syms n
>> limit(((2*n-3)/(3*n-7))^4, inf)

ans =

16/81

>> limit((3*n^3+7*n^2+1)/(4*n^3-8*n+5),n,inf)

ans =

3/4

The last two limits present an uncertainty of the form ∞.0 and ∞0:

>> limit(((n+1)/2)*((n^4+1)/n^5), inf)

ans =

1/2

>> limit(((n+1)/n^2)^(1/n), inf)

ans =

1
18
www.pdfgrip.com
Exercise 2-2

n n
æ 3+n ö æ 2 ö 1 -3 n + 3 1+ n n!
lim ç ÷ , lim ç 1 - ÷ , lim n , lim , lim n
n®¥ -1 + n + n®¥ - n + 1 + n
è ø n®¥
è 3 n ø n ®¥ n n ®¥ n
The first two examples are indeterminate of the form 1∞:

>> limit(((n+3)/(n-1))^n, inf)

ans =

exp(4)

>> limit((1-2/(n+3))^n, inf)

ans =

exp (- 2)

The next two limits are of the form ∞0 and (¥ - ¥) :
¥
>> limit((1/n)^(1/n), inf)

ans =

1

>> maple('limit(((n+1)^(1/3)-n^(1/3))/((n+1)^(1/2)-n^(1/2)),n=infinity)')

ans =

0

The last limit is of the form ¥ ¥ :

>> maple('limit(n!/n^n, n=infinity)')

ans =

0
19
www.pdfgrip.com
2.2 Limits of Functions. Lateral Limits

To calculate the limits of functions one uses the same MATLAB commands as for limits of sequences. For functions,
MATLAB allows you to calculate the limit at a point, and left and right limits (if these limits exist). If a function has a
limit at a point then it necessarily has left and right limits at that point, and they coincide. If the left and right limits
do not coincide then the function does not have a limit at the given point. Below are several exercises which illustrate
how to calculate function limits. Some exercises are accompanied by graphics. The use of graphics is advisable if there
are any doubts concerning the results.
Exercise 2-3
sin é( ax ) ù
2
x
-1 - x x - 2+x x ë û , lim e - 1 .
lim , lim , lim 1 + x , lim
x ®1 -1 + x x ®2 -3 + 1 + 4 x x ®0 x ®0 x 2 x ®0 log (1 + x )
Initially, we have two indeterminates of type 0/0 and one of the form 1∞:

>> syms x
>> limit((x-1)/(x^(1/2)-1),x,1)
2

>> limit((x-(x+2)^(1/2))/((4*x+1)^(1/2)-3),2)

9/8

>> limit((1+x)^(1/x))
exp (1)

The last two are indeterminates of the form 0/0:

>> syms x a, limit(sin(a*x)^2/x^2,x,0)

a^2

>> numeric(limit((exp(1)^x-1)/log(1+x)))

1
20
www.pdfgrip.com
Exercise 2-4
Calculate the following function limits:

|x| x -1
lim , lim x 2 - x - 7 , lim n , lim x e .
x ®1 sin ( x ) x ®3 x ®1 x - 1 x ®0
The first limit is calculated as follows:

>> limit(abs(x)/sin(x),x,0)

ans =

NaN

>> limit(abs(x)/sin(x),x,0,'left')

ans =

-1

>> limit(abs(x)/sin(x),x,0,'right')

ans =

1

As the lateral boundaries are not equal, the function has no limit at x = 0.
If we plot the function (see Figure 2-1), the limits become clear:

>> ezplot (abs (x) /sin (x), [- 1, 1])
Figure 2-1.
21
www.pdfgrip.com
By simple observation, we see that the limit from the right is 1 and the limit from the left is - 1.
For the next two limits we have:

>> limit(abs(x^2-x-7),x,3)

ans =

1

>> limit((x-1)/(x^n-1),x,1)

ans =

1/n

For the last limit we have the following:

>> limit(exp(1)^(1/x),x,0)

ans =

NaN

>> limit(exp(1)^(1/x),x,0,'left')

ans =

0

>> limit(exp(1)^(1/x),x,0,'right')

ans =
INF

Then, there is no limit at x = 0. If we plot the function (see Figure 2-2), we have:

>> ezplot (exp(1)^(1/x), [- 10, 10 - 3, 3])

22
www.pdfgrip.com
Figure 2-2.
We see that the function becomes (positive) infinite at 0 when approaching from the right, and it tends to 0 when
approaching from the left. Thus we conclude that the function has no limit at x = 0.
2.3 Sequences of Functions

MATLAB enables you to analyze the convergence of sequences of functions. MATLAB’s graphics capabilities can
greatly help in understanding the concept of the limit of a sequence of functions and can also aid in the calculation
of these limits.
Exercise 2-5
Calculate the limit function of the following sequences of functions:

fn(x) = x n/n, gn(x) = x n/(1 + x n), hn(x) = x n/(n + x n), kn(x) = sin 2(px) if 1 / (n + 1)£ x £1/n and kn(x) = 0 if (x < 1/(n + 1)
and x > 1/n ). In all cases x varies in [0,1].
First of all, we impose the condition x ∈[0,1] using the syntax:

>> maple('assume(x, RealRange(0,1))');

We then calculate the limits of the function sequences as follows:

>> pretty(sym(maple('limit(x^n/n,n=infinity)')))

0
23
www.pdfgrip.com
>> pretty(sym(maple('limit(x^n/(1+x^n),n=infinity)')))

0
>> pretty(sym(maple('limit(x^n/(n+x^n),n=infinity) ')))

0

>> pretty(sym(maple('limit(piecewise(x>=1/(n+1) and x<=1/n,
sin(Pi/x)^2,x<1/(n+1) and x<1/n,0),n=infinity) ')))

0

Now we support these results with the graphical representations shown in Figures 2-3, 2-4 and 2-5, which in
each case illustrates the convergence of the sequence of functions towards the constant 0 function. The
syntax for the plots is as follows:

>> fplot('[x,x^2/2,x^3/3,x^4/4,x^5/5,x^6/6,x^7/7,x^8/8,x^9/9,x^10/10]',[0,1,-1/2,1])

>> fplot('[x/(1+x),x^2/(1+x^2),x^3/(1+x^3),x^4/(1+x^4),x^5/(1+x^5),x^6/(1+x^6),
x^7/(1+x^7),x^8/(1+x^8),x^9/(1+x^9),x^10/(1+x^10)]',[0,1,-1/2,1])

>> fplot('[x/(1+x),x^2/(2+x^2),x^3/(3+x^3),x^4/(4+x^4),x^5/(5+x^5),x^6/(6+x^6),
x^7/(7+x^7),x^8/(8+x^8),x^9/(9+x^9),x^10/(10+x^10)]',[0,1,-1/2,1])
Figure 2-3.
24
www.pdfgrip.com
Figure 2-4.
Figure 2-5.
Exercise 2-6
Calculate the limit of the following sequences of functions:

fn(x) = (x2 + nx)/n, x∈R
gn(x) = 1 if – n £ x £n and gn(x) = 0 otherwise, x∈R
hn(x) = 2n2x if 0 £ x £1/2n, hn(x) = 2n – 2n2x if 1/2n £ x £n, hn( x ) = 0 if n £ x £1, x ∈ [0,1]
The limits of these sequences of functions are calculated as follows:

>> pretty(sym(maple('limit((x^2+n*x)/n,n=infinity)')))

x
25
www.pdfgrip.com
>> pretty(sym(maple('limit(piecewise(x>=-n and x<=n,1,0),n=infinity)')))

1

>> maple('assume(x, RealRange(0,1))');
>> pretty(sym(maple('limit(piecewise(x>=0 and x<=1/(2*n), 2*n^2*x, x>=1/(2*n) and
1/n>=x, 2*n-2*x*n^2, 1/n<=x and x>=1, 0),n=infinity)')))

0

Now we support these results with the graphical representations shown in Figures 2-6, 2-7 and 2-8, which
illustrate the convergence of the first sequence of functions towards the function f(x) = x, the second
sequence to the function f(x) = 1 and the third sequence to the function f(x) = 0. The syntax entered into
the Matlab/Editor debugger to define the functions via the M-File sub-option of the File menu in the MATLAB
command window is the following:

function f=seq1(x,n)
f=(x^2+n*x)/n;

function g=seq2(x,n)
if x>=-n & x<n
g=1;
else g=0;
end

function h=seq3(x,n)
if x>=0 & x<=1/(2*n)
h=2*n^2*x;
elseif x>1/(2*n) & x<=1/n
h=2*n-2*n^2*x;
elseif x>1/n & x<=1
h = 0;
end

26
www.pdfgrip.com
Figure 2-6.
Figure 2-7.
27
www.pdfgrip.com
Figure 2-8.
We keep the definitions of the sequences of functions in the M-files named seq1.m, seq2.m and seq3.m
respectively. The graphical respresentation of the first 10 functions of each sequence is as follows:

>> fplot ('[seq1(x,1), seq1(x,2), seq1(x,3), seq1(x,4), seq1(x,5),])
([seq1(x,6), seq1(x,7), seq1(x,8), seq1(x,9), seq1(x,10)]', [- 2, 2, - 2, 2])

([seq2(x,6), seq2(x,7), seq2(x,8), seq2(x,9), seq2(x,10)]'[- 10, 10, 0, 3/2])

([seq3(x,6), seq3(x,7), seq3(x,8), seq3(x,9), seq3(x,10)]', [- 0, 1, - 1/2, 10])
2.4 Continuity
A function f is continuous at the point x = a if:
lim f ( x ) = f ( a ) .
x ®a
Otherwise, it is discontinuous at the point. In other words, in order for a function f to be continuous at a it must
be defined at a and the limit of the function at a must be equal to the value of the function at a.
If the limit of f (x) as x tends to a exists but is different to f (a), then f is discontinuous at a, and we say f has an
avoidable discontinuity at a. The discontinuity is resolved by redefining f (a) to coincide with the limit.
If the two lateral limits of f (x) at a exist (whether finite or infinite) but are different, then the discontinuity of f at a
is said to be of the first kind. The difference between the two lateral limits is the jump. If the jump is finite then the
discontinuity is said to be of the first kind with finite jump, otherwise it is of the first kind with infinite jump.
If either of the lateral limits do not exist, then the discontinuity is said to be of the second kind.
28
www.pdfgrip.com
MATLAB provides the following commands relating to continuity:

maple (‘readlib (discont): discont (function, variable)’): Determines the points of
discontinuity of the real valued function given in the specified variable.
maple (‘readlib (iscont): iscont(function,var=a..b)’): Determines when the given
function is continuous on the interval [a, b]. The functional expression can contain
piecewise-defined functions.
maple (‘readlib (iscont): iscont(function,var=a..b, open)’): Determines when the given
given function is continuous on the open interval (a, b).
We illustrate these concepts with several exercises:
Exercise 2-7
Study the continuity of the following functions of a real variable:

sin( x ) æ1ö
f (x) = , g ( x ) = sin ç ÷ .
x èxø

>> syms x a
>> limit(sin(x)/x,x,a)

ans =

sin(a)/a

The function f is continuous at any non-zero point a, so for any such point we have that lim f ( x ) = f ( a ) .
x ®a
The problem arises at the point x = 0, at which the function f is not defined. Therefore, the function is
discontinuous at x = 0, This discontinuity can be avoided by redefining the function at x = 0 with a value
equal to lim f ( x ) .
x ®0

>> limit(sin(x)/x,x,0)

ans =

1

Thus we conclude that the function f ( x ) = sin( x ) / x presents an avoidable discontinuity at x = 0 that is avoided
by defining f (0) = 1. The function is continuous at all non-zero points.
The function g is continuous at any non-zero point a, so lim g ( x ) = g ( a ) .
x ®a

>> limit (sin (1/x), x, a)

ans =

sin(1/a)

29
www.pdfgrip.com
The problem arises at the point x = 0, where the function g is not defined. Therefore, the function is discontinuous
at x = 0. To try to avoid the discontinuity, we calculate lim g ( x ) .
x ®0

>> limit(sin(1/x),x,0)

ans =

-1 .. 1

>> limit(sin(1/x),x,0,'left')

ans =

-1 .. 1

>> limit(sin(1/x),x,0,'right')

ans =
-1 .. 1

We see that the limit does not exist at x = 0 (the limit has to be unique and here the result given is all points in
the interval [− 1,1]), and neither of the lateral limits exist. Thus the function has a discontinuity of the second
kind at x = 0.
MATLAB responded to the calculation of the above limits with the expression "– 1.. 1". This is because the graph
of g (x) presents infinitely many oscillations between – 1 and 1. This is illustrated in Figure 2-9:

>> fplot ('sin (1/x)', [- 1, 1])

Figure 2-9.
30
www.pdfgrip.com
Exercise 2-8
Study the continuity of the following function of a real variable:

1
f (x) = if x ¹ 0 and f ( x ) = 1 if x = 0 .
1+ x e
The only problematic point is x = 0. Now, the function does exist at x = 0 (it has the value 1). We will try to find
the lateral limits as x tends to 0:

>> syms x
>> limit(1/(1+exp(1/x)),x,0,'right')

ans =

0

>> limit(1/(1+exp(1/x)),x,0,'left')

ans =

1

As the lateral boundaries are different, the limit of the function at 0 does not exist. However, the lateral boundaries
are both finite, so the discontinuity at 0 is of the first kind with a finite jump. We illustrate this result in Figure 2-10.

>> fplot('1/(1+exp(1/x))',[-5,5])
Figure 2-10.
31
www.pdfgrip.com
Exercise 2-9
Study the continuity of the following function of a real variable:

f ( x ) = x e if x ¹ 0 and f ( x ) = 1 if x = 0.
The only problematic point is x = 0. The function is defined at x = 0 (it has the value 1). We will try to find the
lateral limits at 0:

>> limit ((exp (x)), x, 0, 'right')

ans =

inf

>> limit((exp(1/x)),x,0, 'left')

ans =

0

As the lateral boundaries are different, the limit of the function as x tends to 0 does not exist. As the right
lateral limit is infinite, the discontinuity of the first kind at x = 0 is an infinite jump. We illustrate this result in
Figure 2-11 above.

>> fplot ('exp (x)', [- 150, 150])
Figure 2-11.
32
www.pdfgrip.com
2.5 Several Variables: Limits and Continuity. Characterization

Theorems
The sequence {(a1n, a2n, …, amn)} of points in m-dimensional space, where n runs through the natural numbers,
has as limit the point (a1, a2, …, am) if, and only if:
lim ( a1n ) = a1 , lim ( a2 n ) = a2 ,..., lim ( amn ) = am .

n®¥ n®¥ n®¥
This characterization theorem allows us to calculate limits of sequences of points in m-dimensional space.
There is another theorem, similar to the above, which characterizes the limits of functions between spaces of
more than one dimension. This theorem enables us to calculate the limits of multivariable functions.
If f : Rn → Rm is a function whose m components are (f1, …, fm). Then it follows that:
lim
x1 ®a1 , x 2 ®a2 ,¼,xn ®an
( f ( x ,x
1 1 2 ,¼, x n ) , f 2 ( x1 , x 2 ,¼, x n ) ,¼, fm ( x1 , x 2 ,¼, x n ) )
= (l1 , l2 ,¼, lm )
if and only if
lim
x1 ®a1 , x 2 ®a2 ,¼,xn ®an
( f (x ,x
1 1 2 ,¼, x n ) ) = l1 ,
lim
x1 ®a1 , x 2 ®a2 ,¼,xn ®an
( f (x ,x
2 1 2 ,¼, x n ) ) = l2 ,
…,
lim
x1 ®a1 , x 2 ®a2 ,¼,xn ®an
( f ( x ,x
m 1 2 ,¼, x n ) ) = lm .
Exercise 2-10
Calculate the limit of the following three-dimensional sequence:
é 1 + n æ 1 ö2 n n ù
lim ê , ç1 + ÷ , ú
n®¥
êë n è n ø 2n - 1 úû

>> pretty(sym(maple('vector([limit((n+1)/n,n=infinity),limit((1+1/n)^(2*n),
n=infinity),limit(n/(2*n-1),n=infinity)])')))

[1-exp (2) 1/2]
33
www.pdfgrip.com
Exercise 2-11
Calculate the limit as n→∞ of the following four-dimensional sequence:
é n 1 n 1 + n2 ù
lim ê n , n , 5n , 2 ú
n®¥
ë 1 + n2 n n û

>> pretty(sym(maple('vector([limit((n/(n^2+1))^(1/n),n=infinity),limit((1/n)^(1/n),
n=infinity),limit((5*n)^(1/n) ,n=infinity),limit((n^2+1)/n^2,n=infinity)])')))

[1, 1, 1, 1]
Exercise 2-12
For the function f : R ® R 2 defined below, find lim f ( x ) :

x ®0
æ sin( x ) x ö
f (x) = ç , 1 + x ÷.
è x ø

>> pretty(sym(maple('vector([limit(sin(x)/x,x=0),limit((1+x)^(1/x),x=0)])')))

[1, exp (1)]
Exercise 2-13
For the function f : R 2 ® R 2 defined below, find lim f (x , y) :

( x ,y )®( 0 ,0 )
é 2(1 - cos( y )) sin( x ) x tan( y ) ù

f (x, y) = ê + , 1+ x - ú.
ë y2 x y û

>> maple ('f:=(x,y) - > sin (x) / x+2 (1-cos (y)) / y ^ 2');

>> maple ('g:=(x,y) - >(1+x) ^ (1/x) - tan (y) / y ');

>> pretty(sym(maple('vector([limit(limit(f(x,y),x=0),y=0),limit(limit(g(x,y),
(((([(x=0), y = 0)])')))
[inf, exp (1) - 1]
34
www.pdfgrip.com
2.6 Iterated and Directional Limits

Given a function f : Rn → R an iterated limit of f at the point (a1, a2, …, an) is the value of the limit (if it exists):
è x2 ®a2
x1 ®a1 (
lim æç lim ¼ lim f ( x1 , x 2 ,¼, x n ) ö÷
xn ®an ø )
or any of the other limits obtained by permuting the order of the component limits.
The directional limit of f at the point (a1, …, an) depends on the direction of the curve h (t) = (h1(t), h2(t), …, hn(t)),
where h(t0) = (a1, a2, …, an), and is defined to be the value:
lim( f (h(t )) = lim f ( x1 , x 2 ,¼, x n )
t ®t 0 ( x1 ,x 2 ,¼,xn )®( a1 ,a2 ,¼,an )
A necessary condition for a function of several variables to have a limit at a point is that all the iterated limits have
the same value (which will be equal to the value of the limit of the function, if it exists).
It can also happen that the directional limit of a function will vary according to the curve used, so that a different
limit exists for different curves, or the limit exists for some curves and not others.
Another necessary condition for a function of several variables to have a limit at a point is that all directional
limits, i.e. the limit for all curves, have the same value.
Therefore, to prove that a function has no limit at a point it is enough to show that either an iterated limit does
not exist or that two iterated limits have a different value, or we can show that a directional limit does not exist or that
two directional limits have different values.
A practical procedure for calculating the limit of a function of several variables is to change from cartesian to
polar coordinates.
Exercise 2-14
Find lim f ( x , y ) for the function f : R 2 → R defined by:

( x , y )®( 0 , 0 )
xy
f (x , y) =
x2 + y2

>> syms x y
>> limit (limit ((x*y) /(x^2+y^2), x, 0), y, 0)

ans =

0

>> limit (limit ((x*y) /(x^2+y^2), y, 0), x, 0)

ans =

0

35
www.pdfgrip.com
Thus the two iterated limits are the same. Next we calculate the directional limits corresponding to the family of
straight lines y = mx:

>> syms m
>> limit((m*x^2)/(x^2+(m^2)*(x^2)),x,0)

ans =

m /(1+m^2)

The directional limits depend on the parameter m, which will be different for different values of m (corresponding
to different straight lines). Thus, we conclude that the function has no limit at (0,0).
Exercise 2-15
Find lim f ( x , y ) for the function f :R 2 → R defined by:

( x , y )®( 0 , 0 )
( y 2 - x 2 )2
f (x , y) =
x2 + y4

>> syms x y
>> limit (limit ((y^2-x^2) ^ 2 /(y^4+x^2), x, 0), y, 0)

ans =

1

>> limit (limit ((y^2-x^2) ^ 2 /(y^4+x^2), y, 0), x, 0)

ans =

0

As the two iterated limits are different, we conclude that the function has no limit at the point (0,0).
36
www.pdfgrip.com
Exercise 2-16

( x , y )®( 0 , 0 )
( y 2 - x )2
f (x , y) =
x2 + y4

>> syms x y m, limit (limit ((y^2-x) ^ 2 /(y^4+x^2), y, 0), x, 0)

ans =

1

>> limit (limit ((y^2-x) ^ 2 /(y^4+x^2), x, 0), y, 0)

ans =

1

Thus the two iterated limits are the same. Next we calculate the directional limits corresponding to the family
of straight lines y = mx:

>> limit(((m*x)^2-x)^2/((m*x)^4+x^2),x,0)

ans =

1

The directional limits corresponding to the family of straight lines y = mx do not depend on m and coincide with
the iterated limits. Next we find the directional limits corresponding to the family of parabolas y ^ 2 = mx:

>> limit(((m*x)-x)^2/((m*x)^2+x^2),x,0)

ans =

(m-1) ^ 2-/(m^2+1)

Thus the directional limits corresponding to this family of parabolas depend on the parameter m, so they are
different. This leads us to conclude that the function has no limit at (0,0).
37
www.pdfgrip.com
Exercise 2-17

( x , y )®( 0 , 0 )
x2y
f (x , y) =
x2 + y2

>> syms x y, limit (limit ((x^2*y) /(x^2+y^2), x, 0), y, 0)

ans =

0

>> limit (limit ((x^2*y) /(x^2+y^2), x, 0), y, 0)

ans =

0

>> limit(((x^2)*(m*x))/(x^2+(m*x)^2),x,0)

ans =

0

>> limit (((m*y) ^ 2) * y / ((m*y) ^ 2 + y ^ 2), y, 0)

ans =

0

We see that the iterated limits and directional limits corresponding to the given family of lines and parabolas
coincide and are all zero. This leads us to suspect that the limit of the function may be zero. To confirm this,
we transform to polar coordinates and find the limit:

>> syms a r, limit (limit (((r^2) * (cos (a) ^ 2) * (r) * (sin (a))) / ((r^2) *
(cos (a) ^ 2)+(r^2) * (sin (a) ^ 2)), r, 0), a, 0)

ans =

0

Therefore we conclude that the limit of the function is zero at the point (0,0).
This is an example where, as a last resort, we had to transform to polar coordinates. In the above examples
we used families of lines and parabolas, but other curves can be used. The change to polar coordinates can
be crucial in determining limits of functions of several variables. As we have seen, there are sufficient criteria
to show that a function has no limit at a point. However, we do not have necessary and sufficient conditions to
ensure the existence of the limit.
38
www.pdfgrip.com
Exercise 2-18
Find lim f ( x , y ) for the function f : R 2 → R defined by:

( x , y )®( 0 , 0 )
( x - 1)2 y 2
f (x , y) =
( x - 1)2 + y 2

>> syms x y m a r
>> limit (limit (y ^ 2 *(x-1) ^ 2 / (y ^ 2 +(x-1) ^ 2), x, 0), y, 0)

ans =

0

>> limit (limit (y ^ 2 *(x-1) ^ 2 / (y ^ 2 +(x-1) ^ 2), y, 0), x, 0)

ans =

0

>> limit((m*x)^2*(x-1)^2/((m*x)^2+(x-1)^2),x,0)

ans =

0

>> limit((m*x)*(x-1)^2/((m*x)+(x-1)^2),x,0)

ans =

0

We see that the iterated and directional limits coincide. We calculate the limit by converting to polar coordinates:

>> limit (limit ((r ^ 2 * sin (a) ^ 2) * (r * cos (a) - 1) ^ 2 / ((r ^ 2 * sin (a) ^ 2) +
(r * cos (a) - 1) ^ 2), r, 1), a, 0)

ans =

0

The limit is zero at the point (1,0). The surface is depicted in Figure 2-12 where we see the function tends to 0 in
a neighborhood of (1,0):

>> [x, y] = meshgrid(0:0.05:2,-2:0.05:2);
>> z=y.^2.*(x-1).^2./(y.^2+(x-1).^2);
>> mesh(x,y,z), view ([- 23, 30])
39
www.pdfgrip.com
Figure 2-12.
2.7 Continuity in Several Variables

A function f : Rn → Rm is said to be continuous at the point (a1, a2, …, an) if:
lim f (a1 , a2 ,¼, an ).

x1 ®a1 ,x 2 ®a2 ,¼,xn ®an

Exercise 2-19
Let the function f : R 2→ R be defined by:

x3 + y3
f (x,y) = .
x2 + y2
Find lim f ( x , y ) and study the continuity of f.

( x , y )®(1, 0 )
The only problematic point is the origin. We will analyze the continuity of the function at the origin by calculating
its limit there.
40
www.pdfgrip.com
>> syms x y m a r
>> limit (limit ((x^3+y^3) /(x^2+y^2), x, 0), y, 0)

ans =

0

>> limit (limit ((x^3+y^3) /(x^2+y^2), y, 0), x, 0)

ans =

0

>> limit((x^3+(m*x)^3)/(x^2+(m*x)^2),x,0)

ans =

0

We see that the iterated and the linear directional limits coincide. We try to calculate the limit by converting to
polar coordinates:

>> maple ('limit (limit (((r * cos (a)) ^ 3 + (r * sin (a)) ^ 3) / ((r * cos (a)) ^ 2 +
(r * sin (a)) ^ 2), r = 0), a = 0)')

ans =

0

We see that the limit at (0,0) coincides with f(0,0), so the function is continuous at (0,0). The graph in Figure 2-13
confirms the continuity.

>> [x, y] = meshgrid(-1:0.007:1,-1:0.007:1);
>> z=(x.^3+y.^3)./(x.^2+y.^2);
>> mesh (x, y, z)
>> view ([50, - 20])
41
www.pdfgrip.com
Figure 2-13.
Exercise 2-20
Define the function f : R 2 ® R by

f ( x , y ) = x 2 + 2 y if ( x , y ) ¹ (1, 2) and f (1, 2 ) = 0
and study its continuity at the point (1,2).

>> limit (limit ((x^2+2*y), x, 1), y, 2).

ans =

5

>> limit (limit ((x^2+2*y), y, 2), x, 1)

ans =

5

We see that if the limit at (1,2) exists, then it should be 5. But the function has the value 0 at the point (1,2). Thus,
the function is not continuous at the point (1,2).
42
www.pdfgrip.com
Exercise 2-21
Consider the function f : R 2 ® R defined by:
f (x, y ) =
[1 - cos( x )]sin( y ) if ( x , y ) ¹ (0 , 0) and f ( 0 , 0 ) = 0.
x +y
3 3
Study the continuity of f at the point (0,0).

>> maple ('limit (limit ((1-cos (x)) * sin (y) /(x^3+y^3), x = 0), y = 0)')

ans =

0

>> maple ('limit (limit ((1-cos (x)) * sin (y) /(x^3+y^3), x = 0), y = 0)')

ans =

0

>> maple ('limit ((1-cos (x)) * sin(m*x) / (x ^ 3, +(m*x) ^ 3), x = 0)')

ans =

1/2 * m /(1 + m^3)

We see that the limit at (0,0) does not exist, as there are different directional limits for different directions. Thus,
the function is not continuous at (0,0). At the rest of the points in the plane, the function is continuous. The plot of
the function in a neighborhood of the origin of radius 0.01 shown in Figure 2-14 shows that around (0,0) there are
branches of the surface that soar to infinity, which causes the non-existence of the limit at the origin.

>> [x,y]=meshgrid(-1/100:0.0009:1/100,-1/100:0.0009:1/100);
>> z=(1-cos(x)).*sin(y)./(x.^3+y.^3);
>> surf (x, y, z)
>> view ([50, - 15])
43
www.pdfgrip.com
Figure 2-14.
44
www.pdfgrip.com
Chapter 3
Numerical Series and

Power Series
This chapter demonstrates the wide range of features that MATLAB offers which can be used to treat numerical series.
These include the determination of the radius of convergence of a power series, summation of convergent series,
alternating series and so on.
3.1 Series. Convergence Criteria

We distinguish between numerical series which have non-negative terms and series which have alternating terms.
Some examples of the more classical criteria to determine whether a series of non-negative terms converges or
diverges will be presented first, then we will go on to analyze alternating series.
3.2 Numerical Series with Non-Negative Terms

Among the most common of the convergence criteria is the ratio test or d'Alembert criterion, which reads as follows:
¥
a(n + 1)
å
n =1
a(n) is convergent if lim
n®¥ a(n )
<1

¥
a(n + 1)
å a(n) is divergent if lim
n =1
n®¥ a(n)
>1
If the limit is 1, we don’t know whether the series converges or diverges.

Another widely used criterion is the Cauchy criterion or root test, which reads as follows:
¥
å a(n) is convergent if lim

n®¥
n
a(n) <1
n =1
å a(n) is divergent if lim

n =1
n®¥
n
a(n) >1
Again, if the limit is 1, we cannot say whether the series diverges or converges.
If a limit of 1 is obtained in both the ratio and root test, we can often use the criterion of Raabe or Duhamel,
which reads as follows:
¥
é æ a(n + 1) ö ù
å a(n) is convergent if lim ên ç 1 - ÷ú > 1
a(n) ø û
n =1
n®¥
ë è
45
www.pdfgrip.com
Chapter 3 ■ Numerical Series and Power Series
¥
é æ a(n + 1) ö ù
å a(n) is divergent if lim ên ç 1 - a(n) ø û
÷ú < 1
n =1
n®¥
ë è
However, if the limit is 1, we still cannot conclude anything about the convergence or divergence of the series.
Another useful criterion is the following:
å a ( n ) and å n=1 2n a ( 2n ) either both diverge or both converge.

¥ ¥
n =1
Other criteria such as Gauss majorization, comparison tests and so on can also be implemented via Maple. We
will see some examples below:
EXERCISE 3-1
Study the convergence and, if possible, find the sum of the following series:
¥
1+ n
å n(2 + n)(3 + n)
n =1
We will apply the ratio test:

>> maple('a:=n -> (n+1)/(n*(n+2)*(n+3)):limit(a(n+1)/a(n), n=infinity);')

ans =

1

We see that the limit is 1, so we can’t yet say anything about the convergence or divergence of the series.
We will apply Raabe’s criterion:

>> maple('a:=n-> (n+1) / (n * (n + 2) * (n)): limit (n * (1-a (n + 1) /a (n)),
n = infinity);')

ans =

2

As the limit is greater than 1, the series converges and will be summable:

>> maple('sum(a(n),n=1..+infinity)')

ans =

17/36

The MATLAB command symsum can be used to sum series, but it is not as strong as the command maple(‘sum’).
The syntax of both is:

syms v, a, b
symsum(S,v,a,b) sums the series with general term S where the variable v ranges between a and b

46
www.pdfgrip.com
symsum(S,v) sums the series with general term S in the variable v

maple('sum(S,v=a..b)') sums the series with general term S where the variable v ranges
between a and b

The sum of the series can also be found as follows:

>> syms n
>> symsum((n+1)/(n*(n+2)*(n+3)),1,Inf)

ans =

17/36

EXERCISE 3-2
¥ ¥ ¥
n nn nn
å2
n =1
n
, å n! , å 3
n =1 n =1
n
n!
We apply the ratio test to the first series:

>> maple('a:=n -> n/2^n: limit(a(n+1)/a(n), n=infinity)')

ans =

1/2

The limit is less than 1, so the series converges. We find its sum:


ans =

2

We apply the ratio test to the second series:

>> maple('a:=n -> n^n/n!: limit(a(n+1)/a(n), n=infinity)')

ans =

exp (1)

As the limit is greater than 1, the series diverges.
47
www.pdfgrip.com
We apply the ratio test to the third series:

>> maple('a:=n -> n^n/((n!)*(3^n)): limit(a(n+1)/a(n), n=infinity)')

ans =

1/3 * exp (1)

The limit is less than 1, so the series is convergent. Trying to find the exact sum, we see that MATLAB does not
solve the problem:


ans =

sum(n^n/n!/(3^n),n = 1 .. inf)
EXERCISE 3-3
¥
3 + 2n ¥
n ¥
( n + p )! , p = parameter
å7
n =1 n (1 + n)
n
, åp
n=1
n
, å
n =1
n
p n!p!
We apply the ratio test to the first series:

>> maple('a:=n -> (2*n+3)/(n*(n+1)*(7^n)): limit(a(n+1)/a(n), n=infinity)')

ans =

1/7

As the limit is less than 1, the series is convergent. We will calculate its sum. The result that is returned will sometimes
be complicated, depending on certain functions that are implemented in MATLAB (in this case hypergeom):


ans =

16/3+30*log(6/7)-13/49*hypergeom([2, 2],[3],1/7)

>> maple('evalf(")')

ans =

.3833972806909634

48
www.pdfgrip.com
Now we apply the ratio test to the second series:

>> maple('a:=n -> n/p^n');
>> maple('limit(a(n+1)/a(n), n=infinity)')

ans =

1/p

Thus, if p > 1, the series converges, and if p < 1, the series diverges. If p = 1, we get the series with general term
n, which diverges. For p greater than 1, we find the sum of the series:


ans =

p /(-1+p) ^ 2

We apply the ratio test to the third series:

>> maple('a:=n ->(n+p)!/((p!)*(n!)*(p^n))');

ans =

limit((n+1+p)!/(n+1)!/(p^(n+1))/(n+p)!*n!*p^n,n = inf)

Here MATLAB has not been able to find the limit, so we will try the Raabe criterion instead:

>> maple('limit(n*(1-a(n+1)/a(n)),n=infinity)')

ans =

limit(n*(1-(n+1+p)!/(n+1)!/(p^(n+1))/(n+p)!*n!*p^n),n = inf)

Again, the limit could not be found. We try to simplify the expression before finding the limit:

>> maple('limit(simplify(a(n+1)/a(n)), n=infinity)')

ans =

1/p

Thus, if p > 1, the series converges, if p < 1, the series diverges and if p = 1, we get the series with general term
n + 1, which diverges.
49
www.pdfgrip.com
EXERCISE 3-4
-n
¥
1 2 ¥ éæ 1 + n ö1+n 1 + n ù
å (1 + ) - n , å êç ÷ - ú
n =1 n n =1 êëè n ø n úû
For the first series we apply the ratio test:

>> maple('a:=n ->(1+1/n)^(-n^2)');
>> maple('simplify(limit(a(n+1)/a(n), n=infinity))')

ans =

exp (- 1)

As the limit is less than 1, the series converges. If we try to find the sum with the commands sum or symsum, we
do not obtain an answer. MATLAB is unable to sum this series, so we can only find an approximate numerical sum:

>> maple ('evalf (sum ((n), n = 1.. + infinity))')

ans =

.8174194332978335

We apply the ratio test to the second series:

>> maple ('a: = n - > (((n + 1) / n) ^ (n + 1) - a (n + 1) /n) ^(-n)');
>> maple ('simplify (limit (to (n + 1) /a (n), n = infinity))')

ans =

1 / (exp (1) - 1)

As the limit is less than 1, the series converges. The exact sum can be found with MATLAB, but we can
approximate the sum using the command evalf:

>> maple('evalf(sum(a(n),n=1..+infinity))')

ans =

1.174585575087587

50
www.pdfgrip.com
EXERCISE 3-5
n2
¥
5 ¥ ¥
æ n 2 + 2n + 1 ö
å , å( n
n - 1) ,
n
å ç 2 ÷
n=1 2n n =1 n =1 è n + n -1 ø
For the first series, we apply the root test:

>> maple('a:=n ->5/(2^n)');
>> maple('limit((a(n))^(1/n), n=infinity)')

ans =

1/2

As the limit is less than 1, the series is convergent:


ans =

5

Now we apply the root test to the second series:

>> maple('a:=n ->(n^(1/n)-1)^n');
>> maple('limit(simplify((a(n))^(1/n)), n=infinity)')

ans =

0

As the limit is less than 1, the series is convergent, but neither the exact nor an accurate approximate sum
are calculable.
Now we apply the root test to the third series:

>> maple('a:=n->((n^2+2*n+1)/(n^2+n-1))^(n^2)');
>> maple('simplify(limit(simplify((a(n))^(1/n)), n=infinity))')

ans =

exp (1)

As the limit is greater than 1, the series diverges.
51
www.pdfgrip.com
EXERCISE 3-6
¥
æ qö
å tan
n =1
n
çp + ÷×
è nø
We apply the root test:

>> maple ('a: = n - > (tan(p+q/n)) ^ n');
>> simple(maple('limit(simplify((a(n))^(1/n)), n=infinity)'))

ans =

sin(p)/cos(p)

Then, for values of p such that tan(p) < 1 the series converges. These values are 0 < p < Pi/4. And for values of p
such that tan(p) > 1 the series diverges. These values are Pi/4 < p < Pi/2. MATLAB does not offer the exact or an
accurate approximate sum of this series.
EXERCISE 3-7
¥ ¥
1 1
å n Log (n) , å n [ Log (n)]
n =1 n =1
p
, p = parameter > 0
å a ( n ) and å 2n a ( 2n ) either both diverge or both converge.

¥ ¥
We apply the criterion that the series n =1 n =1

>> maple ('a: = n - > 1 / (n * log (n))');
>> maple('b:=(2^n)*a(2^n)');
>> pretty(maple('simplify(b)'))

1
---------
n log (2)

As the general term is a constant multiple of the general term of the divergent harmonic series, this series
diverges, and we conclude that the original series diverges.
Let us now apply the same criteria to the second series:

>> maple('a:=n->1/(n*(log(n))^p)');
>> maple('b:=(2^n)*a(2^n)');
>> pretty(maple('simplify(b)'))

52
www.pdfgrip.com
(- p) (- p)
n log(2)

When p < 1, this series dominates the series with general term n –p = 1/np. Thus the initial series also diverges.
When p < 1, this series is dominated by the convergent series with general term n –p = 1/n p. Thus the initial
series also converges.
When p = 1, the series reduces to the series studied above, i.e. it diverges.
MATLAB does not offer the sum of either of the two series of this problem.
EXERCISE 3-8
¥
(n + 1)(n + 2) ¥
1
å , å
(1 + n )
2
n =1 n5 n =1
We will begin by studying the second series. We try to apply the ratio, Raabe and root tests:

>> maple('a:=n->1/(1+n^(1/2))^2');

ans =

1


ans =

1


ans =

1

Thus all the limits are 1. Therefore, at the moment, we cannot conclude anything about the convergence of the series.
We now compare our series with the divergent harmonic series by finding the limit of the quotient of the
respective general terms:

>> maple('limit(a(n)/(1/n),n=infinity)')

ans =

1

53
www.pdfgrip.com
As the limit is greater than zero, the initial series is also divergent.
We will now analyze the first series of the problem directly, comparing it with the convergent series with general
term 1/n3 by examining the limit of the quotient of the general terms:

>> maple('a:=n->(n+1)*(n+2)/n^5');
>> maple('limit(a(n)/(1/(n^3)),n=infinity)')

ans =

1

As the limit is greater than 0, the initial series is also convergent:


ans =

2 * zeta (5) + 1/30 * pi ^ 4 + zeta (3)

Now, we try to approximate the result:

>> maple('evalf(sum(a(n),n=1..+infinity))')

ans =

6.522882114579747

We could have tried to determine whether the first series converges or diverges by using the ratio, root and
Raabe criteria:


ans =

1


ans =

1


ans =

3

The root and ratio tests tells us nothing, but since the limit is greater than 1, the Raabe criterion tells us that
the series converges.
54
www.pdfgrip.com
3.3 Alternating Numerical Series

We now consider numerical series that have alternating positive and negative terms.
A series ∑a(n) is absolutely convergent if the series ∑ |a(n)| is convergent. As the series of moduli is a series of
positive terms, we already know how to analyze it.
Every absolutely convergent series is convergent.
Apart from the criteria described earlier, there are, among others, two classical criteria that allow us to
analyze the nature of alternating series, which will allow us to resolve most convergence problems concerning
alternating series.
The Dirichlet test says that if the sequence of partial sums of ∑a(n) is bounded and {b(n)} is a decreasing
sequence that has limit 0, then the series ∑a(n)b(n) is convergent.
Abel’s test says that if ∑a(n) is convergent and {b(n)} is a monotone convergent sequence, then the
series ∑a(n)b(n) is convergent.
EXERCISE 3-9
Study the convergence of the following series:

¥
(-1)1+n
å 1 + 2n
n =1
2
This is an alternating series. Let us consider the series of moduli and analyze its character:

>> maple ('a: = n - > 1 /(2*n^2+1)');

We apply to this series of positive terms the criteria of comparison of the second kind, comparing with the
convergent series with general term 1/n 2:

>> maple('limit(a(n)/(1/(n^2)),n=infinity)')

ans =

½

As the limit is greater than zero, the given series of positive terms is convergent, so the initial series is absolutely
convergent and, therefore, convergent.
EXERCISE 3-10
Study the convergence of the following series:

¥
(-1)1+n n ¥
(-1)1+n
å
n =1 (1 + n)2
, å
n =1 n
n
Defining a ( n ) = ( -1) and b ( n ) = å a ( n ) has bounded partial sums and {b(n)}
1+n
, we have that
(1 + n )
2
is monotone decreasing with limit 0.

Using the Dirichlet test we conclude that the series is convergent.
55
www.pdfgrip.com
For the second series we similarly proceed as follows.

1
Defining a ( n ) = ( -1) and b ( n ) = , we have that å a (n ) has bounded partial sums and {b ( n )}
1+n
n
is monotone decreasing with limit 0.
Using the Dirichlet test we conclude that the series is convergent.
3.4 Power Series

Given the power series å a ( n )x n, the most pressing issue is to calculate the range of convergence, i.e., the range of
values of x for which the series is convergent.
The most common convergence criteria that we use are the root and ratio tests applied to the series of moduli
(absolute values). If we can show the series is convergent then the original series is absolutely convergent and hence
convergent.
EXERCISE 3-11
Study the range of convergence of the following power series:

¥
42n
å n + 2 ( x - 3)
n =0
n
We apply the ratio test:

>> maple('a:=n->(4^(2*n))*((x-3)^n)/(n+2)');

ans =

16 * x-48

The series will be convergent when | 16 (- 3 + x ) | < 1 :

[solve('16*x-48=1'), solve('16*x-48=-1')]

ans =

49/16 47/16

Thus, the condition 16(-3 + x ) < 1 is equivalent to the following:
47 49 ×
16 < x < 16
We already know that for values of x in the previous interval the series is convergent. Now we need to analyze the
behavior of the series at the end points of the interval. We first consider x = 49/16:

>> maple ('a: = x - > (4 ^(2*n)) * ((x-3) ^ n) / (n + 2)');
>> maple('simplify(a(49/16))')

56
www.pdfgrip.com
ans =

1/(n+2)

We first apply convergence tests for non-negative series to see if any of them determine the convergence or
divergence of the series.
The ratio, Raabe and root tests do not solve the problem. Next we apply the criterion of comparison of the second
kind, comparing the series of the problem with the divergent harmonic series with general term 1/n:

>> maple('a:=n->1/(n+2)');

ans =

1

As the limit is greater than zero, the series is divergent.
We now analyze the behavior of the series at the other end point x = 47/16:

>> maple ('a: = x - > (4 ^(2*n)) * ((x-3) ^ n) / (n + 2)');
>> maple ('simplify (a(47/16))')

ans =

(- 1) ^ n / (n + 2)
¥
(-1)n
We have to analyze the alternating series å n + 2 .
n =1
The series with general term (-1)n has bounded partial sums, and the sequence with general term 1 / (n + 2)
is decreasing toward 0. Then, by the Dirichlet test, the alternating series converges. Therefore the interval of
convergence of the power series is the half-closed interval [47/16, 49/16).
EXERCISE 3-12
Study the range of convergence of the following power series:

¥
1
å (-5)
n =0
n
x 2 n+1
We apply the root test:

>> maple('a:=n->x^(2*n+1)/(-5)^n');

ans =

-1/5 * x ^ 2

57
www.pdfgrip.com
The series is absolutely convergent when |–x ^ 2/5| < 1.

The condition |–x ^ 2/5| < 1 is equivalent to -Sqrt (5) < x < Sqrt (5). Thus, we have determined a possible interval
of convergence of the power series. We will now analyze the end points:

>> maple ('a: = x - > x ^(2*n+1) /(-5) ^ n');
>> [maple('simplify(a(sqrt(5)))'),maple('simplify(a(-sqrt(5)))')]

ans =

(- 1) ^(-n) * 5 ^(1/2) -(-1) ^ n * 5 ^(1/2)

Both series are obviously divergent alternating series. Therefore, the interval of convergence of the power series
is the open interval ( - 5 , 5 ) .
3.5 Formal Power Series

MATLAB implements the package powseries, which can be loaded into memory using the command with(powseries).
This package provides commands which allow you to create, manipulate and perform symbolic calculations with
formal power series. The command maple must first be used. Among these commands are the following:
powcreate(eqn,eqn1,eqn2,..,eqnn): Creates a formal power series, where eqn is an
equation of the form f (n) = expression (n), and where the equations eqni, which are
of the form f(ni) = value, define the initial conditions that will give the values of the
coefficients of the power series
powpoly (polynomial, variable): Creates a formal power series equivalent to the
polynomial given in the specified variable
powadd(ps1,..,psn): Gives the formal power series sum of the specified series
subtract(ps1,ps2): Gives the formal power series subtraction of the specified series
negative (pserie): Gives the additive inverse of the specified power series (i.e. changes
its sign)
multconst (pserie, expression): Multiplies each coefficient of the specified power series
by the given expression
multiply (ps1, ps2): Gives the product of the two specified power series
quotient (ps1, ps2): Gives the quotient of the two specified power series
powsqrt (pserie): Gives the square root of the specified power series
powexp (pserie): Gives the exponential of the specified power series
powlog (pserie): Gives the natural logarithm of the specified power series
powsin (pserie): Gives the sine of the specified power series
powcos (pserie): Gives the cosine of the specified power series
powdiff (pserie): Gives the derivative of the specified power series
powint (pserie): Gives the indefinite integral of the specified power series
58
www.pdfgrip.com
compose (ps1, ps2): Composes the specified power series

reversion (ps1, ps2): Calculates the reversion of the power series ps1 with respect to the
power series ps2 (reversion is the inverse of composition)
inverse (pserie): Calculates the multiplicative inverse of the given series
evalpow (exprseries): Returns the power series obtained by evaluating the expression
exprseries, where the latter is any arithmetic expression involving formal power series,
polynomials, or functions that is accepted by the power series package
tpsform (pserie, variable, n): Transforms the formal power series pserie into a series of
order n in the given variable
tpsform (pserie, variable): Transforms the formal series pserie into a series in the given
variable with order defined by the global variable Order
op (pserie): Returns the internal structure of the formal power series pserie and gives
access to its operands
op (op (pserie) 4): Returns a table that defines all the coefficients of the formal power
series pserie. Any operand can be replaced with subsop.
Here are some examples:

>> maple('with(powseries):powcreate(t(n)=1/n!,t(0)=1):')
>> pretty(simple(sym(maple('tpsform(t, x, 7)'))))

2 3 4 5 6 7
1 + x + 1/2 x + 1/6 x + 1/24 x + 1/120 x + 1/720 x + O(x )

>> maple('powcreate(v(n)=(v(n-1)+v(n-2))/4,v(0)=4,v(1)=2): ')
>> pretty(simple(sym(maple('tpsform(v, x) '))))

2 3 19 4 47 5 6
4 + 2 x + 3/2 x + 7/8 x + -- x + --- x + O(x )
32 128
>> maple('with(powseries): ')
>> maple('t:=powpoly(2*x^5+4*x^4-x+5, x): ')
>> pretty(simple(sym(maple('tpsform(t, s, 5) '))))

4 5
5 - s + 4 s + O(s )

>> pretty(simple(sym(maple('tpsform(t, s, 4)'))))

4
5 - s + O(s )

>> maple('powcreate(t(n)=t(n-1)/n,t(0)=1): ')
>> maple('powcreate(v(n)=v(n-1)/2,v(0)=1): ')
>> maple('s := powadd(t, v): ')
>> pretty(simple(sym(maple('tpsform(s, x, 7) '))))

59
www.pdfgrip.com
2 3 4 19 5 49 6 7
2 + 3/2 x + 3/4 x + 7/24 x + 5/48 x + --- x + ---- x + O(x )
480 2880
>> maple('s := multiply(t, v): ')
>> pretty(simple(sym(maple('tpsform(s, x) '))))

2 19 3 4 109 5 6
1 + 3/2 x + 5/4 x + -- x + 7/16 x + --- x + O(x )
24 480

>> maple('s := powexp(x): ')
>> pretty(simple(sym(maple('tpsform(s, x, 7) '))))

2 3 4 5 6 7
1 + x + 1/2 x + 1/6 x + 1/24 x + 1/120 x + 1/720 x + O(x )

>> maple('t := powexp( exp(x) ): ')
>> pretty(simple(sym(maple('tpsform(t, x, 4) '))))

2 3 4
exp (1) + exp (1) x + exp (1) x + 5/6 exp (1) x + or (x)

>> maple('u := powexp( powdiff( powlog(1+x) ) ): ')
>> pretty(simple(sym(maple('tpsform(u, x, 5) '))))

2 3 73 4 5
exp(1)- exp(1) x + 3/2 exp(1) x - 13/6 exp(1) x + -- exp(1) x + O(x)
24
>> maple('powcreate(t(n)=t(n-1)/n,t(0)=0,t(1)=1): ')
>> maple('powcreate(v(n)=v(n-1)/2,v(0)=0,v(1)=1): ')
>> maple('s := reversion(t,v): ')
>> pretty(simple(sym(maple('tpsform(s,x,11) '))))

3 5 7 9 11
x + 1/12 x + 1/80 x + 1/448 x + 1/2304 x + O(x )

>> maple('ts := compose(t,s): ')
>> pretty(simple(sym(maple('tpsform(ts,x) '))))

2 3 4 5 6
x + 1/2 x + 1/4 x + 1/8 x + 1/16 x + O(x )

>> pretty(simple(sym(maple('tpsform(v,x) '))))

2 3 4 5 6
x + 1/2 x + 1/4 x + 1/8 x + 1/16 x + O(x )

>> maple('powcreate(f(n)=f(n-1)/n,f(0)=1): ')
>> maple('powcreate(g(n)=g(n-1)/2,g(0)=0,g(1)=1): ')
>> maple('powcreate(h(n)=h(n-1)/5,h(0)=1): ')
>> maple('k:=evalpow(f^3+g-quotient(h,f)): ')
>> pretty(simple(sym(maple('tpsform(k,x,5) '))))

60
www.pdfgrip.com
233 2 7273 3 52171 4 5

24/5 x + --- x + ---- x + ----- x + o (x)
50 1500 15000
3.6 Power Series Expansions and Functions

A local approximation of a real function of a real variable at a point replaces the definition of the function in a
neighborhood of the point by a simpler function. The most commonly used local approximations replace an arbitrary
function f(x) by a polynomial p(x) (a truncated power series), so that for any x close to the point, f(x) is close to p(x).
MATLAB implements several commands which can be used to work with local approximations. The command
maple must first be used. The syntax of these commands is presented below:
series(expr,var=a): Returns a truncated power series expansion of expr in a
neighborhood of the point a. The series is truncated in the order specified by the
global variable Order. The expansion can be Taylor, Laurent or other more generalized
series. The expansion may have fractional exponents, in which case the series will be
represented in ordinary sum-of-products form.
series (expr, var): Equivalent to the above with a = 0
series (int (expr, var) var = a): Returns a generalized series expansion of the specified
integral in a neighborhood of the point a
series (leadterm (expr), var = a): Returns the smallest degree term in the generalized
series expansion of expr in a neighborhood of a
series(expr,var=a,n): Returns the generalized series expansion of expr in a
neighborhood of a up to order n. In order to evaluate the expansion at some input, it
must first be converted to a polynomial using the command convert (polynom).
convert (s, polynom): Converts the series s to a polynomial, eliminating the order term
convert (s, ratpoly): Converts the series s to a rational polynomial expression. If s
is a Taylor or Laurent series, then the rational polynomial expression is the Padé
approximation. If s is a Chebyshev series, then the rational polynomial expression is
the Chebishev-Padé approximation.
convert(s,ratpoly,m,n): Converts the series s to a rational polynomial where the
numerator polynomial has degree m and the denominator polynomial has degree n
op (s): Shows the internal structure of the series s and allows access to its operands.
It returns a sequence of 2n operands where n is the number of terms of the series,
including the order. The ith coefficient is the 2i-1th term and the ith exponent is the 2ith
term. If the expansion is in a neighborhood of x = a, op(0,series) extracts the expression
var = a. Any operand can be substituted with subsop.
order (s): Determines the order of the truncated expansion of the series s
Order: = n sets the global variable Order to the value n. By default, the value is 6. This
variable determines the order of all future truncated series expansions.
Order returns the current value of the order at which all truncated series expansions
are given. The default order is 6.
type (expr, series): Determines whether the expression is a series
61
www.pdfgrip.com
type (expr, taylor): Determines whether the expression is a Taylor series

type (expr, laurent): Determines whether the expression is a Laurent series

>> pretty(simple(sym(maple('series(x/(1-x-x^2), x=0) '))))

2 3 4 5 6
x + x + 2 x + 3 x + 5 x + O(x )

>> pretty(simple(sym(maple('series(exp(x)/x, x=0, 8 ) '))))

-1 2 3 4 5 6 7
x + 1 + 1/2 x + 1/6 x + 1/24 x + 1/120 x + 1/720x + 1/5040 x + O(x)

>> pretty(simple(sym(maple('series(GAMMA(x), x=0, 2 ) '))))

-1 2 2 2
x - gamma + (1/12 Pi + 1/2 gamma ) x + O(x )

>> pretty(simple(sym(maple('series(x^3/(x^4+4*x-5),x=infinity) '))))

4 5 1
1/x - ---- + ---- + O(----)
4 5 7
x x x

>> pretty(simple(sym(maple('int(exp(x^3), x ) : series(", x=0) '))))

4 7
x + 1/4 x + O(x )

>> pretty(simple(sym(maple('series(x^x, x=0, 3) '))))

2 2 3
1 + ln(x) x + 1/2 ln(x) x + O(x )

>> pretty(simple(sym(maple('convert(" ,polynom) '))))

2 2
1 + ln(x) x + 1/2 ln(x) x

>> pretty(simple(sym(maple('series(exp(x), x) '))))

2 3 4 5 6
1 + x + 1/2 x + 1/6 x + 1/24 x + 1/120 x + O(x )

62
www.pdfgrip.com
>> pretty(simple(sym(maple('convert(", ratpoly) '))))

2 3
1 + 3/5 x + 3/20 x + 1/60 x
-----------------------------
2
1 - 2/5 x + 1/20 x
3.7 Taylor, Laurent, Padé and Chebyshev Expansions

Taylor’s theorem gives us a way of approximating a function f(x) by a polynomial at a point x = x0. The series that is
truncated to give such a polynomial is called the Taylor series of the function f(x) at the point x = x0. The Taylor series
of f(x) at x = 0 is called the MacLaurin series of f(x).
Taylor’s theorem says that if f(x) is an n + 1times differentiable function defined on an interval that contains the
point a, then f(x) can be approximated in a neighborhood of a by the following polynomial:
n
f ( k ) (a ) f (n+1) (z )
f (x) = å ( x - a )k + ( x - a )n+1
k =0 k! (n + 1)!
where z is a number between a and x.

The above expression is the Taylor series expansion of f(x) at the point x = a. The last term of the expression is the
remainder term, and represents a measure of the error of the approximation.
The Laurent expansion allows a finite number of negative powers in the series expansion of a function.
The Padé expansion approximates the function as a ratio of polynomials whose numerator and denominator
have fixed degrees m and n.
The Chebyshev expansion approximates the function using Chebyshev polynomials. The Chebyshev-Padé
expansion mixes both types.
When you use the command “series”, without further specification, MATLAB uses by default the type of expansion
most suitable in each case. There are commands that allow you to specify which type of expansion should be used (all
require the prior use of the maple command). The syntax is as follows:
taylor(expr,var=a): Returns the Taylor series expansion of the expression expr at the
point a where the order is determined by the global variable Order
taylor (expr, variable): Returns the Taylor series expansion of the expression expr in a
neighborhood of the origin (i.e. the MacLaurin expansion)
taylor(expr,var=a,n): Returns the Taylor expansion of expr at a up to order n.
taylor(int,var=a): Returns the Taylor expansion of the indefinite integral int in a
neighborhood of a
taylor(leadterm(expr),var=a): Returns the lowest degree term in the Taylor series
expansion of expr in a neighborhood of a
readlib (mtaylor): Loads into the memory a package that allows you to work with
multivariate Taylor expansions
mtaylor (expression, [var1,..., varn] or mtaylor(expression,{var1,...,varn})):
Returns the multivariate Taylor expansion of expr given in the specified variables up to
the order determined by the global variable Order
mtaylor(expression,[var1,...,varn],n):
63
www.pdfgrip.com
Returns the multivariate Taylor expansion of the specified expression up to order n

mtaylor(expression,[var1,...,varn],n,[w1,...,wn]):
Returns the multivariate Taylor expansion of the given expression by assigning weights
w1 to wn to the variables. The default weight is 1. A weight of 2 will halve the order of
the corresponding variable to which the expansion is given.
readlib (coeftayl): Loads into the memory a package that allows you to find coefficients
of Taylor expansions
coeftayl(expr,variable=a,n): Returns the coefficient of xn in the Taylor expansion of
expr in a neighborhood of the point a.
coeftayl(expr,[var1,..,varm]=[a1,...,an],[n1,...,nm]):
Returns the coefficient of the term (var1-a1)n1 ... (varm-am)nm in the Taylor expansion
of expr in a neighborhood of the point (a1,..., am)
readlib (poisson): Loads into the memory a package that allows you to work with
multivariate Taylor series in Poisson form
poisson(expression,[var1,...,varn]) or poisson(expr,{var1,...,varn}):
Returns the multivariate Taylor expansion of the expression in Poisson form. The
Poisson form combines the terms of the series in terms of sines and cosines, forming
the Fourier canonical form. The order of the expansion is specified by the global
variable Order.
poisson(expr,[var1,...,varn],n):
Returns the multivariate Taylor expansion of expr in Poisson form up to order n
poisson(expr,[var1,...,varn],n,[w1,...,wn]):
Returns the multivariate Taylor expansion of order n of expr in Poisson form using the
specified weights
chebyshev(expr,var=a..b):
Returns the Chebyshev expansion of an analytical expression expr in the interval [a, b].
The result is a sum based on orthogonal Chebyshev polynomials in t. The orthopoly
package needs to be loaded via the command with(orthopoly,t) before applying the
Chebyshev command.
chebyshev(expr,var): Returns the Chebyshev expansion of an analytical expression
expr in the interval [- 1,1]
chebyshev(expr,var=a..b, error):
Returns the Chebyshev expansion with the specified tolerance, which by default
is 10 ^(-Digits)
asymt (expr, var): Returns the asymptotic expansion (var -> infinity) of the given
expression expr up to the order specified by the global variable Order
asymt (expr, var, n): Returns the asymptotic expansion (var -> infinity) of the given
expression expr up to order n
with (numapprox): Loads into the memory a package that allows you to work with
Padé, Padé-Chebyshev and Laurent expansions
64
www.pdfgrip.com
chebdeg (series): Returns the degree of the specified Chebyshev series

chebmult (s1, s2): Returns the product of the two specified Chebychev series in the
same variable
chebpade(expr,var=a..b,[n,d]): Finds the Chebyshev-Padé approximation of the given
expression expr in the specified variable, in the interval [a, b], with numerator and
denominator polynomials of degrees n and d respectively
chebpade(expr,var,[n,d]): Returns the Chebyshev-Padé approximation of the given
expression expr in the specified variable, in the interval [- 1,1], with numerator and
denominator polynomials of degrees n and d respectively
chebpade(expr,var=a.. b, [n, 0]) or chebpade(expr,var=a..b, n): Simply returns the
Chebyshev polynomial of degree n in [a, b]
chebsort (series): Orders the terms of the given Chebyshev series
pade(expr,var=a,[n,d]): Returns the Padé approximation of the given expression expr
in the specified variable at the point a, with numerator and denominator polynomials
of degrees n and d respectively
pade(expr,var,[n,d]): Returns the Padé approximation of the given expression expr in
the specified variable, at 0, with numerator and denominator polynomials of degrees
n and d respectively
pade(expr,var=a,[n,0]) or pade(expr,var=a,n):
Simply returns the Taylor series expansion up to order n of the expression at the given
point
laurent(expr,var=a,n):
Returns the Laurent series of the given expression expr at the given point up to order n
laurent(expr,var,n): Returns the Laurent series of the given expression expr at the point
0 up to order n
readlib (eulermac) loads into the memory a package whose commands allow you to
find Euler-MacLaurin expansions
eulermac(expr,var,n): Finds the Euler-MacLaurin expansion of the given expression up
to degree n in the given variable
eulermac (expr, var): Finds the Euler-MacLaurin expansion of the given expression up
to the degree specified by the global variable Order
EXERCISE 3-13
Find the Taylor series up to 13th order of sinh (x) at the point x = 0. Also find the Taylor expansion of 1 /(1+x) up to
10th order at the point x = 0. Find the corresponding Laurent and Euler-McLaurin series. Find the Chebychev-Padé
approximation of degree (4,3) in the interval [0,1].

>> pretty(simple(sym(maple('taylor(sinh(x),x=0,13) '))))

65
www.pdfgrip.com
3 5 7 9 11 13
x + 1/6 x + 1/120 x + 1/5040 x + 1/362880 x + 1/39916800 x + O(x )

>> pretty(simple(sym(maple('taylor(1/(1+x),x=0,10) '))))

2 3 4 5 6 7 8 9 10
1 - x + x - x + x - x + x - x + x - x + O(x )

>> maple('with(numapprox): ')
>> pretty(simple(sym(maple('laurent(sinh(x),x=0,13) '))))

3 5 7 9 11 13
x + 1/6 x + 1/120 x + 1/5040 x + 1/362880 x + 1/39916800 x + O(x )

>> pretty(simple(sym(maple('laurent(1/(1+x),x=0,10) '))))

2 3 4 5 6 7 8 9 10
1 - x + x - x + x - x + x - x + x - x + O(x )

Note that the Laurent and Taylor series agree in this case.

>> maple('readlib(eulermac) : ')
>> pretty(simple(sym(maple('eulermac(sinh(x),x,13) '))))

314416268077
------------ cosh(x) - 1/2 sinh(x) + O(cosh(x))
290594304000

>> pretty(simple(sym(maple('eulermac(1/(1+x),x,10) '))))

1 1 1
ln(1 + x) - 1/2 ----- - 1/12 -------- + 1/120 --------
1 + x 2 4
(1 + x) (1 + x)

1 1 1 1
- 1/252 -------- + 1/240 -------- - 1/132 --------- + O(---------)
6 8 10 12
(1 + x) (1 + x) (1 + x) (1 + x)

>> maple(' with(numapprox): ')
>> pretty(simple(sym(maple(' with(orthopoly,T):chebpade(sinh(x),x=0..1,[4,3]) '))))

2
(.03672350156 + .9734745062 x - .02546155828 (2. x - 1.)
3 4
+ .009531362262 (2. x - 1.) - .001730582994 (2. x - 1.) ) /
2
(1.157361890 - .3056450150 x - .009078765811 (2. x - 1.)
3
+ .001929812884 (2. x - 1.) )
66
www.pdfgrip.com
>> pretty(simple(sym(maple('with(orthopoly,T):chebpade(1/(1+x),x=0..1,[4,3]) '))))

.6666666665
---------------------------
.6666666666 + .6666666668 x

EXERCISE 3-14
Find the Laurent expansion of 1 / (x sin (x)) at the point x = 0 up to degree 10. Also find the Padé approximation of
degree (3,4) at the same point and the Taylor expansion of log (x) at the point x = 2 up to degree 5.

>> pretty(simple(sym(maple('with(numapprox):laurent(1/(x*sin(x)),x=0,10) '))))

-2 2 31 4 127 6 7
x + 1/6 + 7/360 x + ----- x + ------ x + O(x )
15120 604800

>> pretty(simple(sym(maple('with(numapprox):pade(1/(x*sin(x)),x=0,[3,4]) '))))

2
-60 - 3 x
------------
4 2
7 x - 60 x

>> pretty(simple(sym(maple('taylor(log(x),x=2,4) '))))

2 3 4
ln(2) + 1/2 (x - 2) - 1/8 (x - 2) + 1/24 (x - 2) + O((x - 2) )
EXERCISE 3-15
Find the Taylor series at the origin, both in its normal form and in its Poisson form, up to order 5, of the function:
2
f(x, y) = e x + y
>> maple('readlib(mtaylor): ')

>> pretty(simple(sym(maple('mtaylor(exp(x+y^2),[x,y],5) '))))

2 2 2 3 4 2 2 4
1 + x + y + 1/2 x + x y + 1/6 x + 1/2 y + 1/2 y x + 1/24 x

>> maple('readlib(poisson): ')
>> pretty(simple(sym(maple('poisson(exp(x+y^2),[x,y],5) '))))

67
www.pdfgrip.com
2 2 2 3 4 2 2 4
1 + x + y + 1/2 x + x y + 1/6 x + 1/2 y + 1/2 y x + 1/24 x
We see that the two forms coincide.
EXERCISE 3-16
Find the Taylor expansion of the function 1 /(2−x) at the point x = 1 up to order 10. Also find the Taylor expansion
of sin(x) at pi/2 up to order 8, and the same for the function log (x) at the point x = 2 up to degree 5.
>> pretty(simple(sym(maple('taylor(1/(2-x),x=1,10)'))))

1 + x - 1 + (x - 1) + (x - 1) + (x - 1) + (x - 1) + (x - 1) + (x - 1) +

8 9 10
(x - 1) + (x - 1) + O((x - 1) )

>> pretty(simple(sym(maple('taylor(sin(x),x=pi/2,8)'))))
2 4 6 8
1 - 1/2(x - 1/2 pi) + 1/24(x - 1/2 pi) - 1/720(x - 1/2 pi) + O((x - 1/2 pi) )

>> pretty(simple(sym(maple('taylor(log(x),x=2,5)'))))

2 3 4 5
log(2) + 1/2 (x - 2) - 1/8 (x - 2) + 1/24 (x - 2) - 1/64 (x - 2) + O((x - 2) )
68
www.pdfgrip.com
Chapter 4
Derivatives and Applications. One

and Several Variables
4.1 The Concept of the Derivative

The derivative of a real function at a point is the instantaneous rate of change of that function in a neighborhood of
the point; i.e., it is a measure of how the dependent variable changes as a result of a small change in the independent
variable.
Geometrically, the derivative of a function at a point represents the gradient of the tangent to the function at the
point. The origin of the idea of the derivative comes precisely from the attempt to draw the tangent line at a given
point on a curve.
A function f(x) defined in a neighborhood of a point x = a is differentiable at a if the following limit exists:
f (a + h ) - f (a )
lim = f ¢(a )
h®0 h
The value of the limit, if it exists, is denoted by f ' (a), and is called the derivative of the function f at the point a.
If f is differentiable at every point of its domain, it is simply said to be differentiable.
The continuity of a function is a necessary condition for its differentiablity, and all differentiable functions are
continuous.
MATLAB can make use of the Maple library is differentiable, whose commands allow you to analyze the
differentiability of a function, and calculate the points where it is not differentiable. The syntax is as follows:
readlib (isdifferentiable): isdifferentiable (expression, var, n, name)
This determines whether the given expression is differentiable to order n in the variable var.
The expression can contain piecewise-defined functions. The optional variable name stores the
sequence of points where the expression is not differentiable.
Exercise 4-1
Study the differentiability of the function:

æ1ö
f ( x ) = x sin ç ÷ if x ¹ 0 and f(x) = 0 if x = 0.
èxø
69
www.pdfgrip.com
Chapter 4 ■ Derivatives and Applications. One and Several Variables
First of all, we study the differentiability at the point x = 0:

>> syms h, limit((h*sin(1/h) - 0)/h,h,0)

ans =

-1 .. 1

1
The limit does not exist, since in any neighborhood of 0 the function sin takes both values - 1 and 1 infinitely
x
many times. Hence MATLAB responds to the calculation of this limit with -1..1.
Let’s see what happens at a non-zero point x = a:

>> maple ('f: = x - > x * sin (1)');
>> maple('limit((f(a+h)-f(a))/h,h=a)')

ans =

2 * sin(1/2/a) - sin (1/a)

Thus we have found the value of the derivative for any non-zero point x = a. The function is plotted in Figure 4-1:

>> fplot ('x * sin (x)', [-1/10.1/10])
Figure 4-1.
70
www.pdfgrip.com
You can also determine whether the function is differentiable using the following:

>> maple('readlib(isdifferentiable):')
>> maple('isdifferentiable(piecewise(x=0,0,x*sin(1/x)),x,0)')

ans =

false
Exercise 4-2
Study the differentiability of the function:
x3
f (x) = if x ¹ 0 and f (x) = 0 if x = 0.
|x|

>> maple('f:=x->x^3/abs(x)');
>> maple('limit((f(0+h)- 0)/h,h=0,left)')

ans =

0

>> maple('limit((f(0+h)- 0)/h,h=0,right)')

ans =

0

Thus we see that the derivative at zero exists and has the value zero. The function is plotted in Figure 4-2, from
which we see that it appears to be differentiable at all points of its domain:

>> fplot('x^3/abs(x)',[-1/10,1/10])
71
www.pdfgrip.com
Figure 4-2.

4.2 Calculating Derivatives

MATLAB provides several commands that allow you to calculate derivatives. We have:
diff('f', 'x') is the derivative of the function f with respect to x
syms x, diff(f,x) is the derivative of the function f with respect to x
diff('f', 'x', n) is the nth derivative of the function f with respect to x
syms x, diff(f, x, n) is the nth derivative of the function f with respect to x
The following commands can also be used for differentiation, but to use them it is necessary to prepend the
command maple.
D (function) returns the derivative of the given expression or function
D (function) (num) returns the value of the derivative of the function at the point num
D (var -> expression) returns the partial derivative of the expression with respect to the
given variable
D [n] (function) returns the partial derivative of the function with respect to its nth
argument
D [n1, n2] (function) returns the partial derivative of the function, first with respect to
its n1th argument and then with respect to its n2th argument
(D@@n)(function) returns the nth derivative of the function
72
www.pdfgrip.com
(D@@n)(f)(num) returns the value of the nth derivative of f at the point num
diff (expr, var) returns the derivative of expr with respect to the specified variable
diff (expr, var$ n) returns the derivative of order n of expr with respect to the specified
variable
diff(expr,var1,var2,...,varn):
Returns the partial derivative of expr with respect to the variables var1 to varn in this
order
Diff(expr,var) returns the inert derivative of expr with respect to the variable var
Diff(expr,var1,..,varn) returns the inert partial derivative of expr with respect to var1 to
varn in this order
op (Diff (expr, var)) shows the internal structure of Diff and allows access to its
operands
op (n, Diff(expr,var)) extracts the nth operand (the first is expr, the second is var,
the 3rd, 4th,... are the variables). You can replace any operand with subsop.
convert (expr, D) converts an expression of the form diff (f (x), x) to the form D (f) (x)
convert (expr, diff) converts an expression of the form D (f) (x) to diff (f (x), x)
implicitdiff (equation, vardep, varind) returns the derivative (dy/dx) of the implicit
function specified by the given equation
implicitdiff (equation, vardep, varind, n) returns the derivative of order n (d yn / dxn) of
the implicit function defined by the given equation
implicitdiff(equation,vardep,varind1,...,varindn) returns the derivative (dny/dx1...dxn)
of the implicit function defined by the given equation
implicitdiff (expression, vardep, varindp) returns the derivative of the implicit function
defined by the equation expression = 0
readlib(isdiffentiable):isdifferentiable(expression,var,n): This determines if the
expression is differentiable to order n in the variable var. The expression can contain
piecewise defined functions.
Exercise 4-3
Calculate the derivative with respect to x of the following functions:
log (sin ( 2 x ) ) , x
tan( x )
,
4 x2 1
3 x2 + 2 (
, log x + x2 + 1 .)
>> pretty(simple(diff('log(sin(2*x))','x')))

2
--------
tan(2 x)
>> pretty(simple(diff('x^tanx','x')))

73
www.pdfgrip.com
tanx
x tanx
------------
x

>> pretty(simple(diff('(4/3)*sqrt((x^2-1)/(x^2+2))','x')))

x
4 -----------------------
2 1/2 2 3/2
(x - 1) (x + 2)

>> pretty(simple(diff('log(x+(x^2+1)^(1/2))','x')))

1
------------
2 1/2
(x + 1)
Exercise 4-4
Calculate the nth derivative of the following functions:
1 x 2 1+ x
, e ,
x 1- x
>> f='1/x';
>> [diff(f),diff(f,2),diff(f,3),diff(f,4),diff(f,5)]

ans =

-1/x ^ 2 2/x ^ 3 - 6/x ^ 4 24/x ^ 5 - 120/x ^ 6

We see from the pattern here that the nth derivative is given by (-1n)+1n ! .
n
x
>> f='exp(x/2)';
>> [diff(f),diff(f,2),diff(f,3),diff(f,4),diff(f,5)]

ans =

1/2*exp(1/2*x) 1/4*exp(1/2*x) 1/8*exp(1/2*x) 1/16*exp(1/2*x) 1/32*exp(1/2*x)
x /2
Thus the nth derivative is e n .
2
>> f='(1+x)/(1-x)';
>> [simple(diff(f)),simple(diff(f,2)),simple(diff(f,3)),simple(diff(f,4))]

74
www.pdfgrip.com
ans =

2 /(-1+x) ^ 2-4 /(-1+x) ^ 3 12 /(-1+x) ^ 4-48 /(-1+x) ^ 5

2 (n !)
Thus, the nth derivative is given by .
(1 - x )n+1
4.3 Tangents, Asymptotes, Concavity, Convexity, Maxima and

Minima, Inflection Points and Growth
If f is a function which is differentiable at x0, then f ' (x0) is the slope of the tangent line to the curve y = f(x) at the point
(x0, f(x0)). The equation of the tangent will be y − f(x0) = f ' (x0) (x − x0).
The horizontal asymptotes of the curve y = f(x) are limit tangents, as x0 → ∞, which are horizontal. They are
defined by the equation y = lim f ( x 0 ).
x 0 ®¥
The vertical asymptotes of the curve y = f(x) are limit tangents, as f(x0) → ∞, which are vertical. They are defined by
the equation x = x0, where x0 is a value such that lim f ( x ) = ¥ .
x ®x0
The oblique asymptotes to the curve y = f(x) at the point x = x0 have the equation y = mx + n, where
y
m = lim and n = lim( y - mx ) .
x ®¥ x x ®¥
If f is a function for which f ' (x0) and f '' (x0) both exist, then, if f ' (x0) = 0 and f '' (x0) < 0, the function f has a local
maximum at the point (x0, f (x0)).
If f is a function for which f ' (x0) and f '' (x0) both exist, then, if f ' (x0) = 0 and f '' (x0) > 0, the function f has a local
minimum at the point (x0, f (x0)).
If f is a function for which f ' (x0), f '' (x0) and f ''' (x0) exist, then, if f ' (x0) = 0 and f '' (x0) = 0 and f ''' (x0) ≠ 0, the
function f has a turning point at the point (x0, f(x0)).
If f is differentiable, then the values of x for which the function f is increasing are those for which f ' (x) is greater
than zero.
If f is differentiable, then the values of x for which the function f is decreasing are those for which f ' (x) is less than zero.
If f is twice differentiable, then the values of x for which the function f is concave are those for which f '' (x) is
greater than zero.
If f is twice differentiable, then the values of x for which the function f is convex are those for which f '' (x) is less
than zero.
Exercise 4-5
Find the equation of the tangent to the curve:

f(x) = 2x 3 + 3x 2 − 12x + 7 at x = − 1.
x2 - x - 4
Also find the x for which the tangents to the curve g ( x ) = are horizontal and vertical. Find their
x -1
asymptotes.

>> f ='2 * x ^ 3 + 3 * x ^ 2-12 * x + 7';
>> g = diff (f)

g =

6*x^2+6*x-12

>> subs(g,-1)

75
www.pdfgrip.com
ans =

-12

>> subs(f,-1)

ans =

20

Thus the slope of the tangent line at the point x = −1 is − 12, and the function at x = −1 has the value 20.
Therefore the equation of the tangent to the curve at the point (− 1, 20) is:
y - 20 = - 12 ( x - (- 1))
We plot the curve and its tangent on the same axes (see Figure 4-3):

>> fplot('[2*x^3+3*x^2-12*x+7, 20-12*(x - (-1))]',[-4,4])

Figure 4-3.
76
www.pdfgrip.com
To calculate the horizontal tangent to the curve y = f(x), we find the values x0 for which f ' (x0) = 0. The equation of
this tangent will then be y = f(x0).
To calculate the vertical tangent to the curve y = f(x), we find the values x0 for which f ' (x0) = ¥. The equation of
this tangent will then be x = x0.

>> g ='(x^2-x+4) /(x-1)'
>> solve(diff(g))

ans =

[3]
[-1]

>> subs(g,3)

ans =

5

>> subs(g,-1)

ans =

-3

The two horizontal tangents will have equations:
y = g '[−1] (x + 1) −3, that is, y = −3
y = g '[3] (x − 3) + 5, that is, y = 5
The horizontal tangents are not asymptotes because the corresponding values of x0 are finite (they are - 1 and 3).
We now find the vertical tangents. To do this, we calculate the values of x for which g'(x) = ¥ (i.e. values for which
the denominator of g ' is zero, but does not cancel with the numerator):

>> solve('x-1')

ans =

1

Therefore, the vertical tangent has equation x = 1.
For x = 1, the value of g(x) is infinite (see below), so the vertical tangent is a vertical asymptote.

subs(g,1)

Error, division by zero

77
www.pdfgrip.com
As lim g ( x ) = ¥ , there are no horizontal asymptotes.

x ®¥
Now let us see if there are any oblique asymptotes:

>> syms x,limit(((x^2-x+4)/(x-1))/x,x,inf)

ans =

1

>> syms x,limit(((x^2-x+4)/(x-1) - x)/x,x,inf)

ans =

0

Thus, the diagonal y = x is an asymptote.
We plot the curve with its asymptotes and tangents:
Using the default command plot, we represent on the same axes (see Figure 4-4) the curve whose equation is
g (x) = (x2 − x + 4) /(x − 1), the horizontal tangents with equations y = − 3 and y = 5, and the oblique asymptote
with equation y = x:

>> fplot('[(x^2-x+4)/(x-1),-3,5,x]',[-10,10,-20,20])
Figure 4-4.

78
www.pdfgrip.com
Exercise 4-6
Find the asymptotes, maxima, minima, inflection points, intervals of growth and decrease and intervals of
concavity and convexity for the function:
x3
f (x) =
x -1
2
>> f='x^3/(x^2-1)'

f =

x^3/(x^2-1)

>> syms x, limit(x^3/(x^2-1),x,inf)

ans =

NaN

Therefore, there are no horizontal asymptotes. To see if there are any vertical asymptotes, let us look at the
values of x that make y infinite:

>> solve('x^2-1')

ans =

[1]
[-1]

Thus the vertical asymptotes are the lines x = 1 and x = −1. Now let us see if there are any oblique asymptotes:

>> limit(x^3/(x^2-1)/x,x,inf)

ans =

1

>> limit(x^3/(x^2-1)-x,x,inf)

ans =

0

The line y = x is an oblique asymptote. Now we shall find the maxima and minima, inflection points and intervals
of concavity:

>> solve (diff (f))

79
www.pdfgrip.com
ans =

[ 0]
[ 0]
[3 ^(1/2)]
[^(1/2) - 3]

The first derivative vanishes at x = 0, x = ÷3 and x = −÷3. These include maximum and minimum candidates.
To verify if they are maxima or minima, we find the value of the second derivative at these points:

>> [numeric(subs(diff(f,2),0)),numeric(subs(diff(f,2),sqrt(3))),
numeric(subs(diff(f,2),-sqrt(3)))]

ans =

0 2.5981 - 2.5981

Therefore, at x = –÷3 there is a maximum and at x =÷3 there is a minimum. At x = 0 we know nothing:

>> [numeric (subs (f, sqrt (3))), numeric (subs (f, - sqrt (3)))]

ans =

2.5981 - 2.5981

Therefore, the maximum point is (−÷3, −2.5981) and the minimum point is (Ö3, 2.5981).
We will now analyze the points of inflection:

>> solve(diff(f,2))

[ 0]
[ i*3^(1/2)]
[-i*3^(1/2)]

The only possible turning point occurs at x = 0, and as f (0) = 0, the possible turning point is (0,0):

>> subs (diff(f,3), 0)

ans =

-6

As the third derivative at x = 0 is non-zero, the origin really is a turning point:

>> pretty(simple(diff(f)))

2 2
x (x - 3)
------------
2 2
(x - 1)

80
www.pdfgrip.com
The curve is increasing when y' > 0, i.e., in the intervals (-¥,-÷ 3) and (÷ 3,¥).
The curve is decreasing when y ' < 0, i.e., in the intervals
(-÷ 3,-1), (-1,0), (0,1) and (1, ÷ 3).

>> pretty(simple(diff(f,2)))

2
x (x + 3)
2 ------------
2 3
(x - 1)

The curve will be concave when y" > 0, i.e., in the intervals (-1,0) and (1, ¥ ).
The curve is convex when y" < 0, i.e. in the intervals (0,1) and (- ¥ , - 1).
The curve has a horizontal tangent at the three points at which the first derivative is zero. The equations of the
horizontal tangents are y = 0, y = -2.5981 and y = 2.5981.
The curve has vertical tangents at the points that make the first deriviative infinite. These are x = 1 and x = −1.
Therefore, the vertical tangents coincide with the two vertical asymptotes.
We plot the curve together with its asymptotes (see Figure 4-5):

>> fplot('[x^3/(x^2-1),x]',[-5,5,-5,5])
Figure 4-5.
81
www.pdfgrip.com
We can also represent the curve, its asymptotes and its horizontal and vertical tangents in the same graph
(Figure 4-6):

>> fplot('[x^3/(x^2-1),x,2.5981,-2.5981]',[-5,5,-5,5])
Figure 4-6.

4.4 Applications to Practical Problems

Many practical problems can be expressed in terms of a function which we need to maximize or minimize, where
these maxima and minima have some applied interpretation. Let’s see some examples:
Exercise 4-7
Write a positive number a as the sum of two summands in such a way that the sum of their cubes is minimal.
Let x be one of the summands. The other will be a − x. We require the sum x 3 + (a − x)3 to be minimized.

>> f='x^3+(a-x)^3'

f =

x^3+(a-x)^3

82
www.pdfgrip.com
>> solve(diff(f,'x'))

ans =

1/2 * a

The possible maximum or minimum is at x = a/2. We use the second derivative to verify that it is indeed a minimum:

>> subs(diff(f,'x',2),'a/2')

ans =

3 * a

As a > 0 (by hypothesis), 3a > 0, which means x = a/2 is a minimum.
Therefore the two summands must be equal to a/2.
Exercise 4-8
Suppose you want to purchase a rectangular plot of 1600 square meters and then fence it. Knowing that the
fence costs 200 cents per meter, what dimensions must the plot of land have to ensure that the fencing is most
economical?
If the surface area is 1600 square feet and one of its dimensions, unknown, is x, and the other will be 1600/x.
The perimeter of the rectangle is p (x) = 2 x + 2 (1600/x), and the cost is given by f(x) = 200 p(x):

>> f ='200 * (2 * x + 2 *(1600/x))'

f =

200 * (2 * x + 2 *(1600/x))

This is the function to minimize:

>> solve (diff (f))

ans =

[40]
[-40]

The possible maxima and minima are presented for x = − 40 and x = 40. To determine their nature, we look at
the second derivative at these points.

>> [subs (diff(f,2), 40), subs (diff(f,2), - 40)]

ans =

20 -20

83
www.pdfgrip.com
x = 40 is a minimum, and x = − 40 is a maximum. Thus, one of the sides of the rectangular field must be 40
meters, and the other will measure 1600/40 = 40 meters. Therefore the optimal rectangle is a square with sides
of 40 meters.
4.5 Partial Derivatives

As we know, MATLAB can differentiate functions with the command diff, but we can also make use of the
differentiation operator D, which calculates derivatives of functions, as well as partial derivatives for functions that
depend on more than one variable.
If you define a function maple('f:=x->...'), to calculate its derivative with respect to the single variable x you can
use the command maple ('D (f )'). To find the value of the derivative at the point x = x0 , i.e., the slope of the function at
the point x = x0, we use the command maple ('D (f )(x0)').
We also know that to calculate the derivative of order n of the function f with respect to x, we can use, apart from
the command diff, @ operators and D, writing maple ('(D@@n) (f )').
If a function depends on several variables (for simplicity, we shall restrict ourselves to a function of two variables,
x and y) we can write it as maple ('g:=(x,y) - >...). In this case, the partial derivatives of the function g can be calculated
in the following way:
g’x : maple(‘D[1](g)’);
g’y : maple(‘D[2](g)’);
g”xx : maple(‘D[1,1](g)’);
g”xy : maple(‘D[1,2](g)’);
g”yy : maple(‘D[2,2](g)’);
givxxyy : maple(‘D[1,1,2,2](g)’).
This notation is generalizable to n variables.
Exercise 4-9
Given the function f (x,y) = sin(xy) + cos(xy 2), calculate:

¶f /¶ x, ¶ f /¶y, ¶2f /¶x 2, ¶2f /¶y 2, ¶2f /¶x¶y, ¶2f /¶y¶x and ¶4f /¶2x ¶ 2y.

>> maple ('f:=(x,y) - > sin(x*y) + cos(x*y^2)');

>> pretty(sym(maple('D[1](f)')))
2 2
(x,y) -> cos(x y) y - sin(x y ) y

>> pretty(sym(maple('D[2](f)')))
2
(x,y) -> cos(x y) x - 2 sin(x y ) x y

>> pretty(sym(maple('D[1,1](f)')))
2 2 4
(x,y) -> - sin(x y) y - cos(x y ) y

84
www.pdfgrip.com
2 2 2 2 2
(x,y) -> - sin(x y) x - 4 cos(x y ) x y - 2 sin(x y ) x

2 3 2
(x,y) -> - sin(x y) y x + cos(x y) - 2 cos(x y ) y x - 2 sin(x y ) y

2 3 2
(x,y) -> - sin(x y) y x + cos(x y) - 2 cos(x y ) y x - 2 sin(x y ) y

>> pretty(sym(maple('D[1,1,2,2](f)')))
2 2 2 6 2
(x,y) -> sin(x y) y x - 4 cos(x y) y x - 2 sin(x y) + 4 cos(x y ) y x
2 4 2 2
+ 18 sin(x y ) y x - 12 cos(x y ) y
4.6 Implicit Differentiation

The differential operator D allows you to find derivatives of functions defined implicitly by an equation. Alternatively,
we can use the command implicitdiff, which we introduced at the beginning of the chapter.
Exercise 4-10
Calculate the implicit derivative of y with respect to x, where x and y satisfy the following equation:
x3 + 3 x2y3 + 5xy + 6y3 = 8
First, we will use the operator D to differentiate the equation:

>> pretty (sym (maple ('D(x^3+3*x^2*y^3+5*x*y+6*y^3=8)')))

2 3 2 2 2
3 D(x) x + 6 D(x) x y + 9 x D(y) y + 5 D(x) y + 5 x D(y) + 18 D(y) y = 0

Now we replace D (x) with 1 and solve for D (y):

>> pretty (sym (maple ('subs (D (x) = 1, D(x^3+3*x^2*y^3+5*x*y+6*y^3=8))')))

2 3 2 2 2
3 x + 6 x y + 9 x D(y) y + 5 y + 5 x D(y) + 18 D(y) y = 0

85
www.pdfgrip.com
>> pretty (sym (maple ('solve (3 * x ^ 2 + 6 * x * y ^ 3 + 9 * x ^ 2 * D (y) * y ^ 2 + 5 *

y + 5 * x * D (y) + + 18 * D (y) * y ^ 2 = 0, D (y))')))

2 3
3 x + 6 x y + 5 y
- ----------------------
2 2 2
9 x y + 5 x + 18 y

A faster and easier way to find implicit first and second order derivatives is to use the command implicitdiff in the
following way:

>> pretty (sym (maple ('implicitdiff(x^3+3*x^2*y^3+5*x*y+6*y^3=8,y,x)')))

2 3
3 x + 6 x y + 5 y
- ---------------------
2 2 2
9 x y + 5 x + 18 y

>> pretty (sym (maple ('implicitdiff(x^3+3*x^2*y^3+5*x*y+6*y^3=8,y,x$2)')))

2 3 3 4 5 3 3 3 4
-2 y (810 x y + 648 x y + 135 y x + 81 x y + 972 x y - 810 x y

4 6 2 2 2 6 4 6 4
- 405 x y - 525 x y - 324 x y - 540 x y - 125 x + 972 y + 162 x

6 / 2 2 4 3 2 6 3 4
+ 81 x ) / (1350 x y + 4860 x y + 125 x + 8748 x y + 4860 x y
/

6 6 5 4 4 6 4 2 6
+ 729 x y + 1215 x y + 4374 x y + 675 x y + 5832 y )
Exercise 4-11
Find the tangents to the ellipse 2 x2 - 2xy + y2 + x + 2y + 1 = 0 at the point x =−3/2.

The slope of the tangent is determined by y' (-3/2):

>> pretty(sym(maple('subs(D(x)=1,D(2*x^2 -2*x*y+y^2+x+2*y+1=0 ))')))

4 x - 2 y - 2 x D(y) + 2 D(y) y + 1 + 2 D(y) = 0

>> pretty(sym(maple('solve(4*x-2*y-2*x*D(y)+2*D(y)*y+1+2*D(y) = 0,D(y))')))

4 x - 2 y + 1
1/2 -------------
x - y - 1

86
www.pdfgrip.com
Alternatively, the implicit derivative can be calculated as follows:

>> pretty(sym(maple('implicitdiff(2*x^2-2*x*y+y^2+x+2*y+1,y,x )')))
4 x - 2 y + 1
1/2 -------------
x - y - 1

Now, to find the values y for which (x, y) is a point on the ellipse with x = − 3/2, we substitute x = − 3/2 into the
equation of the ellipse and solve for y.

>> maple ('solve (subs(x=-3/2,2*x^2-2*x*y+y^2+x+2*y+1=0), y)')

ans =

-1, - 4

The points of contact of the two tangents will be (-3/2,-1) and (-3/2, - 4). Now let’s calculate the slopes of the two
possible tangents:

(((>> maple('subs(x=-3/2,y=-1,-(4*x-2*y+1)/(-2*x+2*y+2))')

ans =

1

(((>> maple('subs(x=-3/2,y=-4,-(4*x-2*y+1)/(-2*x+2*y+2))')

ans =

1

The equations of the two tangents to the curve at x = − 3/2 are then:
y + 1 = x + 3/2 and y + 4 = x + 3/2.
We graphically represent the ellipse, the tangents and the coordinate axes (see Figure 4-7). As the curve is given
in implicit coordinates, we first modify it. To do this, we try to complete the squares in order to find a trigonometric
parameterization that eliminates possible fractional powers arising when trying to separate a variable.
87
www.pdfgrip.com
Figure 4-7.
MATLAB has the command maple(‘completesquare’), implemented in the student library, which allows you to
complete the square of an expression with respect to a specified variable.
We complete the square of the defining expression of the ellipse with respect to the variable y as follows:

>> pretty(maple('completesquare(2*x^2-2*x*y+y^2+x+2*y+1,y)'))

2 2
(y - x + 1) + x + 3 x

Now, it is trivial to separate the variable y for the given equation: y = x - 1 + - x 2 - 3 x . But we have an
expression within a square root, which we rewrite as follows:

>> pretty(maple('completesquare(-x^2-3*x,x)'))

2
- (x + 3/2) + 9/4

We obtain the equation of the curve

2
æ 3ö
y = x -1+ -ç x + ÷ + 9 / 4
è 2ø
3
which is easily parametrized by putting x + 3/2 = 3/2 sin (t ), so that x = * ( sin (t ) - 1) , y = 3 / 2 * (sin (t ) + cos (t )) - 5 / 2.
2
Now we can create the graph of the curve, the asymptotes and the coordinate axes (the OY axis is represented in
parametric form as x = 0, y = t ).
88
www.pdfgrip.com
>> fplot('[x+3/2-1,x+3/2-4,0]',[-4,2,-6,2]);
>> hold on;
>> t=(0:.1:2*pi);
>> x=3/2*(sin(t)-1);
>> y=3/2*(sin(t)+cos(t))-5/2;
>> plot(x,y);
>> t=(-6:.1:2);
>> x=zeros(size(t));
>> y=t;
>> plot(x,y)

4.7 Differentiation of Functions of Several Variables

We now present some of the features that MATLAB offers to treat differentiation of functions of several variables.
Given a function f: R2→ R, the partial derivative of f with respect to the variable x at the point (a, b) is defined as
follows:
¶f f (a + h , b ) - f (a , b )
(a , b) = lim
¶x h®0 h
Similarly, the partial derivative of f with respect to the variable y at the point (a, b) is defined as:
¶f f (a , b + h ) - f (a , b )
(a , b) = lim
¶y h®0 h
Generally speaking, we can define the partial derivative with respect to any variable for a function of n variables.
Given the function f: Rn→ R, the partial derivative of f with respect to the variable xi (i = 1, 2,..., n) at the point
(a1, a2, ..., an) is defined as follows:
¶f f (a1 , a2 ,..., ai + h ,..., an ) - f (a1 , a2 ,..., an )

(a1 , a2 ,..., an ) = lim
¶xi h®0 h
The function f is differentiable if all partial derivatives with respect to xi (i = 1, 2,..., n) exist and are continuous.
All differentiable functions are continuous, and if a function is not continuous, it cannot be differentiable.
The directional derivative of the function f with respect to the vector v = (v1,v2,...,vn) is defined as the following
scalar product:
æ¶f ¶f ¶f ö
(Df )v = ç , ,..., ÷ · (v1 ,v2 ,...,vn ) = ( Ñf ) · v
¶
è 1 x ¶ x 2 ¶ xn ø
æ¶f ¶f ¶f ö
Ñf = ç , ,..., ÷ is called the gradient vector of f.
è ¶ x1 ¶ x 2 ¶ xn ø
The directional derivative of the function f with respect to the vector v = (dx1,dx2,...,dxn) is called the total
differential of f. Its value is:
æ¶f ¶f ¶f ö
Df = ç dx1 + dx 2 + ... + dx n ÷
¶
è 1 x ¶ x 2 ¶ x n ø
89
www.pdfgrip.com
At the beginning of the chapter (under the heading “Calculating derivatives”) we defined the functions that
MATLAB implements to find partial derivatives. Using these functions, we have the following:

∂f
diff (f (x, y, z...), x) = —— = maple ('D [1] (f)(x,y,z,...)') = maple ('diff (f(x,y,z,...), x)')
∂x

∂nf n)
diff (f(x,y,...), x, n) = —— = maple('D[1,1,...1] (f)(x,y,...)') = maple ('diff (f(x,y,...), x$n)')
∂xn

∂f
diff(f(x1,x2,...),xj) = ———— maple('D[ j ](f)(x1,x2,...)') = maple(‘diff(f(x1,x2,..), xj)’)
∂xj

∂nf n)
diff(f(x1,...),xj,n)= ———— = maple('D[j, j,...j](f)(x1,...)') = maple(‘diff(f(x1,...), xj $n)’)
∂xjn

∂∂∂
maple ('diff (f(x,y,z,...), x, y, z,...)') = ———————— ...f = maple('D[1,2,3..])(f)(x,y,z,...)')
∂x ∂y ∂z

∂2f
maple ('diff (f(x,y), x, y)') = ——————— = maple ('D [1,2] (f)(x,y)')
∂x∂y

∂(k)f
maple(‘diff(f(x1,...),xn,xm,xp)’)= ————————— = maple(‘D[n,m,p,...](f)(x1,x2,...)’)
∂xn∂xm∂xp ... (k = number of variables)

Exercise 4-12
Given the real-valued two-variable function f defined by:

xy
f (x,y) = if x2 + y2 ¹ 0 and f(x, y) = 0 if x2 + y2 = 0
x + y2
2
calculate the partial derivatives of f at the origin. Study the differentiability of f.

>> maple ('f:=(x,y) - >(x*y) /(x^2+y^2)');

To find ¶f/¶x and ¶f/¶y at the point (0,0), we directly apply the definition of a partial derivative at a point:

>> maple('(f(h,0)-0)/h')

ans =

0

>> maple('(f(0,k)-0)/k')

90
www.pdfgrip.com
ans =

0

The limits of the two previous expressions as h  0 and k  0, respectively, are both zero.
We see that the two partial derivatives at the origin are the same and have value zero. But the function is not
differentiable at the origin, because it is not continuous at (0,0), since there is no limit as (x, y ) tends to (0,0 ):

>> maple('limit((m*x)^2/(x^2+(m*x)^2),x=0)')

ans =

m ^ 2 /(1+m^2)

The limit does not exist at (0,0), because approaching the origin along different straight lines y = mx yields
different results depending on the parameter m.
Exercise 4-13
Study the differentiability of the following function:

2 xy
f (x,y) = if (x, y ) ¹ (0, 0) and f (0, 0) = 0.
x2 + y2
The function is differentiable if it has continuous partial derivatives at every point. We will consider any point other
than the origin and calculate the partial derivative with respect to the variable x :

>> maple('f:=(x,y)->(2*x*y)/(x^2+y^2)^(1/2)');
>> pretty(simple(sym(maple('D[1](f)(x,y)'))))

3
y
2 --------------
2 2 3/2
(x + y )

Now, let’s see if this partial derivative is continuous at the origin:

>> maple('limit(2*(m*x)^3/(x^2+(m*x)^2)^(3/2),x=0)')

ans =

2*m^3/(1+m^2)^(3/2)

The limit does not exist at (0,0), because approaching the origin along different straight lines y = mx yields
different results depending on the parameter m. Therefore, the partial derivative is not continuous at the origin.
We conclude that the function is not differentiable.
91
www.pdfgrip.com
However, the function is continuous, since the only problematic point is the origin, and the limit of the function at
the origin is 0 = f (0,0):

>> maple('limit(limit(f(x,y),x=0),y=0)')

ans =

0

>> maple('limit(limit(f(x,y),y=0),x=0)')

ans =

0

>> maple('limit(f(x,(m*x)),x=0)')

ans =

0

>> maple('limit(f(x,(m*x)^(1/2)),x=0)')

ans =

0

>> maple('limit(f((r*cos(a)),(r*sin(a))),r=0)')

ans =

0

The iterated limits and the directional limits are all zero, and by changing the function to polar coordinates, the
limit at the origin turns out to be zero, which coincides with the value of the function at the origin.
This is therefore an example of a non-differentiable continuous function.
Exercise 4-14
Calculate the total differential of the following functions:

3 + x 3 − 2 x 2 y , e ax cos (by ).

>> maple('f:=(x,y)->x^3-2*x^2*y+3');

>> pretty(simplify(sym('D(f(x,y))')))
2 2
3 D(x) x - 4 D(x) x y - 2 x D(y)

92
www.pdfgrip.com
Grouping the terms together, we obtain the expression of the total differential:

>> pretty(sym(maple('collect(D(f(x,y)),D(x))')))

2 2
(3 x - 4 x y) D(x) - 2 x D(y)

>> maple('g:=(x,y)->exp(a*x)*cos(b*y)');
>> pretty(sym(maple('dotprod([D[1](g)(x,y),D[2](g)(x,y)],[dx,dy])')))

a exp(a x) cos(b y) dx - exp(a x) sin(b y) b dy
Exercise 4-15
1
then ¶ f2 + ¶ f2 + ¶ f2 = 0.
2 2 2
Verify that if f ( x , y , z ) =
x +y +z
2 2 2
¶x ¶y ¶z
>> f='1/(x^2+y^2+z^2)^(1/2)';
>> simplify(symop(diff(f,'x',2),'+',diff(f,'y',2),'+',diff(f,'z',2)))

ans =

0

Another path would be the following:

>> maple('f:=x->1/(x^2+y^2+z^2)^(1/2)');
>> simplify(sym('diff(f(x,y,z),x$2)+diff(f(x,y,z),y$2)+diff(f(x,y,z),z$2)'))

ans =

0

Alternatively, another way to do the same is as follows:

>> maple('simplify(D[1,1](f)(x,y,z)+D[2,2](f)(x,y,z)+D[3,3](f)(x,y,z))')

ans =

0
Exercise 4-16
Find the directional derivative of the function:
1
f (x , y ,z) =
x + y2 + z2
2
at the point (2,1,1), with respect to the vector v = (1,1,0).

93
www.pdfgrip.com
>> maple('f:=(x,y,z)->1/(x^2+y^2+z^2)^(1/2)');
>> maple('dotprod([D[1](f)(2,1,1),D[2](f)(2,1,1),D[3](f)(2,1,1)],[1,1,0])')

ans =

-1/12*6^(1/2)

Exercise 4-17
Given the function:

x2 +y2
-
f (x, y) = e 8 éëcos2 ( x ) + sin 2 ( y )ùû
calculate:
¶ f ¶ f ¶2f ¶2f ¶2f ¶3f ¶4f ¶5f
, , , , , , and
¶x ¶y ¶x¶y ¶x ¶y ¶x¶y ¶x ¶y
2 2 2 2 2
¶ x3 ¶ y2
and find its value at the point (3p, 6p).

>> maple('f:=(x,y)->exp(-(x^2+y^2)/8)*(cos(x)^2+sin(y)^2)');
>> pretty(factor(sym('D[1](f)(x,y)')))

2 2 2 2
- 1/4 exp(- 1/8 x - 1/8 y ) (x cos(x) + x sin(y) + 8 cos(x) sin(x))

>> pretty(factor(sym('D[1](f)(pi/3,pi/6)')))
2 1/2
- 1/24 exp(- 5/288 pi ) (pi + 12 3 )

If an approximation is required, we can find it as follows:

>> numeric('D[1](f)(pi/3,pi/6)')

ans =

-0.8399

>> pretty(factor(sym('D[2](f)(x,y)')))

2 2 2 2
- 1/4 exp(- 1/8 x - 1/8 y ) (y cos(x) + y sin(y) - 8 sin(y) cos(y))

>> pretty(factor(sym('D[2](f)(pi/3,pi/6)')))

2 1/2
- 1/48 exp(- 5/288 pi ) (pi - 24 3 )

>> numeric('D[2](f)(pi/3,pi/6)')

94
www.pdfgrip.com
ans =

0.6745

>> pretty(factor(sym('D[1,2](f)(x,y)')))

2 2
1/16 exp(- 1/8 x - 1/8 y )

2 2
(y x cos(x) + y x sin(y) + 8 y cos(x) sin(x) - 8 x sin(y) cos(y))

>> pretty(factor(sym('D[1,2](f)(pi/3,pi/6)')))

2 1/2
1/576 pi exp(- 5/288 pi ) (pi - 12 3 )

>> numeric('D[1,2](f)(pi/3,pi/6)')

ans =

-0.0811


2 2
1/16 exp(- 1/8 x - 1/8 y )

2 2
(y x cos(x) + y x sin(y) + 8 y cos(x) sin(x) - 8 x sin(y) cos(y))


2 1/2
1/576 pi exp(- 5/288 pi ) (pi - 12 3 )

>> numeric('D[2,1](f)(pi/3,pi/6)')

ans =

-0.0811


2 2 2 2 2 2 2 2
1/16 exp(- 1/8 x - 1/8 y ) (- 36 cos(x) - 4 sin(y) + x cos(x) + x sin(y)

2
+ 16 x cos(x) sin(x) + 32 sin(x) )

95
www.pdfgrip.com

2 2 1/2
1/288 exp(- 5/288 pi ) (252 + pi + 24 pi 3 )

>> numeric('D[1,1](f)(pi/3,pi/6)')

ans =

1.1481


2 2 2 2 2 2 2 2
1/16 exp(- 1/8 x - 1/8 y ) (- 4 cos(x) - 36 sin(y) + y cos(x) + y sin(y)

2
- 16 y sin(y) cos(y) + 32 cos(y) )


2 2 1/2
- 1/1152 exp(- 5/288 pi ) (- 1008 - pi + 48 pi 3 )

>> numeric('D[2,2](f)(pi/3,pi/6)')

ans =

0.5534

>> pretty(factor(sym('D[1,1,1,2](f)(x,y)')))

2 2 2 2
1/256 exp(- 1/8 x - 1/8 y ) (- 108 y x cos(x) - 12 y x sin(y)

3 2 3 2
- 608 y cos(x) sin(x) + y x cos(x) + y x sin(y)

2 2
+ 24 y x cos(x) sin(x) + 96 y x sin(x) + 96 x sin(y) cos(y)

3
- 8 x sin(y) cos(y))

>> pretty(factor(sym('D[1,1,1,2](f)(pi/3,pi/6)')))

2 3 1/2 2 1/2
1/82944 pi exp(- 5/288 pi ) (pi + 12 3 pi + 756 pi - 5616 3 )

>> numeric('D[1,1,1,2](f)(pi/3,pi/6)')

96
www.pdfgrip.com
ans =

-0.2271

>> pretty(factor(sym('D[1,1,2,2](f)(x,y)')))

2 2 2 2 2 2
1/256 exp(- 1/8 x - 1/8 y ) (144 cos(x) + 144 sin(y) - 4 x cos(x)

2 2 2 2 2
- 36 x sin(y) - 64 x cos(x) sin(x) - 128 sin(x) - 36 y cos(x)

2 2 2 2 2 2 2 2 2
- 4 y sin(y) + y x cos(x) + y x sin(y) + 16 y x cos(x) sin(x)

2 2 2 2
+ 32 y sin(x) + 64 y sin(y) cos(y) - 16 y x sin(y) cos(y) - 128 cos(y)

2 2
+ 32 x cos(y) )

>> pretty(factor(sym('D[1,1,2,2](f)(pi/3,pi/6)')))

1/165888*
2 2 1/2 4 3 1/2
exp(- 5/288 pi ) (- 77760 + 1260 pi - 1728 pi 3 + pi - 24 pi 3 )

>> numeric('D[1,1,2,2](f)(pi/3,pi/6)')

ans =

-0.3856

>> pretty(factor(sym('D[1,1,1,2,2](f)(x,y)')))

2 2 2
- 1/1024 exp(- 1/8 x - 1/8 y ) (2432 cos(x) sin(x) + 432 x cos(x)

2 2 3 2 3 2
+ 432 x sin(y) - 384 x sin(x) - 4 x cos(x) - 36 x sin(y)

2 2 2 2
+ 192 y x sin(y) cos(y) + 96 y x sin(x) - 108 y x cos(x)

2 3 2 2 3 2 2 2 2 2
+ y x cos(x) + y x sin(y) + 24 y x cos(x) sin(x) - 12 y x sin(y)

2 2 3 2
- 608 y cos(x) sin(x) - 96 x cos(x) sin(x) + 32 x cos(y)

2 3
- 384 x cos(y) - 16 y x sin(y) cos(y))

97
www.pdfgrip.com
>> pretty(factor(sym('D[1,1,1,2,2](f)(pi/3,pi/6)')))

2
1/1990656 exp(- 5/288 pi )*

1/2 3 2 1/2 4 1/2 5
(- 1181952 3 + 233280 pi - 1764 pi + 8208 pi 3 + 12 pi 3 - pi )

>> numeric('D[1,1,1,2,2](f)(pi/3,pi/6)')

ans =

-0.5193

The advantage of using the operator D instead of the command diff is that it is much faster at directly calculating
the value of the derivative at a point.
4.8 Maxima and Minima of Functions of Several Variables

A function f : Rn→R, which maps the point (x1, x2,..., xn)ÎR to f (x1,x2,...,xn)ÎR, has an extreme point at (a1, a2, ..., an) if the
æ¶f ¶f ¶f ö
gradient vector Ñf = ç , ,..., ÷ is zero at (a1, a2, ..., an).
è ¶ x1 ¶ x 2 ¶ xn ø
By setting all the first order partial derivatives equal to zero and solving the resulting system, we can find the
possible maxima and minima.
To determine the nature of the extreme point, it is necessary to construct the Hessian matrix, which is defined
as follows:
é ¶2f ¶2f ¶2f ù

ê ........
ê ¶ x1
2
¶ x1 ¶ x 2 ¶ x1 ¶ x n úú
ê ¶2f ¶2f ¶2f ú
ê ......... ú
H = ê ¶ x1 ¶ x 2 ¶ x 22 ¶ x2 ¶ xn ú
ê ........ ........ ........ ........ ú
ê ú
ê ¶2f ¶2f ¶2f ú
ê ........ ú
ë ¶ x1 ¶ x n ¶ x2 ¶ xn ¶ x n2 û
First, suppose that the determinant of H is non-zero at the point (a1, a2, ..., an). In this case, we say that the point is
non-degenerate and, in addition, we can determine the nature of the extreme point via the following conditions:
If the Hessian matrix at the point (a1, a2, ..., an) is positive definite, then the function has a minimum at that point.
If the Hessian matrix at the point (a1, a2, ..., an) is negative definite, then the function has a maximum at that point.
In any other case, the function has a saddle point at (a1, a2, ..., an).
If the determinant of H is zero at the point (a1, a2, ..., an), we say that the point is degenerate.
MATLAB includes a specific function which calculates the determinant of the Hessian of a function of several
variables. Its syntax is:

maple ('hessian(function, [x1,x2,..., xn])')

98
www.pdfgrip.com
MATLAB also includes a function which calculates the gradient of a function of several variables. Its syntax is as
follows:

maple ('grad(function, [x1,x2,..., xn])')
Exercise 4-18
Find and classify the extreme points of the function:
f ( x , y ) = -120 x 3 - 30 x 4 + 18 x 5 + 5 x 6 + 30 xy 2 .
First, we find the extreme points, setting the partial derivatives (the components of the gradient vector of f ) to
zero and solving the resulting system:

>> maple('f:=(x,y)->-120*x^3-30*x^4+18*x^5+5*x^6+30*x* y^2 ');
>> maple('solve({diff(f(x,y),x)=0,diff(f(x,y),y)=0},{x,y})')

ans =

{x = 0, y = 0}, {y = 0, x = 2}, {y = 0, x = -3}, {y = 0, x = -2}

Thus, the extreme points are: (− 2,0), (2,0), (0,0) and (− 3.0).
Alternatively, we can find the gradient vector and solve the system obtained by setting it equal to zero:

>> pretty(sym(maple('grad(f(x,y),[x,y])')))

[-360*x^2-120*x^3+90*x^4+30*x^5+30*y^2, 60*x*y]

>> maple('solve({-360*x^2-120*x^3+90*x^4+30*x^5+30*y^2=0, 60*x*y=0},{x,y})')

ans =

{x = 0, y = 0}, {y = 0, x = 2}, {y = 0, x = -3}, {y = 0, x = -2}

To classify the extreme points, we construct the Hessian matrix, and calculate its value at each point:

>> pretty(sym(maple('hessian(-120*x^3-30*x^4+18*x^5+5*x^6+30*x* y^2,[x,y])')))

[-720*x-360*x^2+360*x^3+150*x^4, 60*y]
[ 60*y, 60*x]

You can also find the Hessian matrix in the following way:

>> pretty(sym(maple('array([[D[1,1](f)(x,y),D[1,2](f)(x,y)],[D[1,2](f)(x,y),
D[2,2](f)(x,y)]])')))

[-720*x-360*x^2+360*x^3+150*x^4, 60*y]
[ 60*y, 60*x]

99
www.pdfgrip.com
But what we need is the Hessian matrix M as a function of x and y so we can evaluate it at certain points:

>> maple ('m:=(x,y) - >
hessian(-120*x^3-30*x^4+18*x^5+5*x^6+30*x* y^2,[x,y])');
>> pretty (sym (maple ('subs (x = 0, y = 0, M(x,y))')))

[0, 0]
[0, 0]

The origin turns out to be a degenerate point, as the determinant of the Hessian matrix is zero at (0,0).

>> pretty (sym (maple ('subs (x =-2, y = 0, M(x,y))')))

[- 480, 0]
[0, - 120]

>> maple ('H: = subs (x =-2, y = 0, M(x,y))');
>> maple('det(H)')

ans =

57600

>> maple('eigenvals(H)')

ans =

-480, - 120

The Hessian matrix at the point (− 2,0) has non-zero determinant, and is also negative definite, because all its
eigenvalues are negative. Therefore, the point (− 2,0) is a maximum of the function.

>> H = sym (maple ('subs (x = 2, y = 0, M(x,y))'))

H =

[2400, 0]
[ 0, 120]

>> determ(H)

ans =

288000

>> eigensys(H)

ans =

[120].
[2400].

100
www.pdfgrip.com
The Hessian matrix at the point (2,0) has non-zero determinant, and is furthermore positive definite, because all
its eigenvalues are positive. Therefore, the point (2,0) is a minimum of the function.

>> H = sym (maple ('subs (x = 3, y = 0, M(x,y))'))

H =

[1350, 0]
[ 0, -180]

>> determ(H)

ans =

-243000

>> eigensys(H)

ans =

[-180.]
[1350]

The Hessian matrix at the point (− 3,0) has non-zero determinant, and, in addition, is neither positive definite nor
negative definite, because its eigenvalues are not all positive or all negative. Therefore (− 3.0) is a saddle point of
the function.
Exercise 4-19
f ( x , y , z ) = x 2 + xy + y 2 + z 2
First, we set the partial derivatives (the components of the gradient vector of f) to zero and solve the resulting
system:

>> maple ('f:=(x,y) - > x ^ 2 + y ^ 2 + z ^ 2 + x * y ');
>> maple ('solve ({diff (f (x, y, z), x) = 0, diff (f (x, y, z), y) = 0, diff (f (x, y, z),
z)}, {x, y, z})')

ans =

{z = 0, y = 0, x = 0}

The single extreme point is the origin (0,0,0). We determine what kind of extreme point it is. To do this, we
calculate the Hessian matrix and express it as a function of x, y and z :

>> maple ('h:=(x,y,z) - > hessian (f (x, y, z), [x, y, z])');
>> M = maple ('H (x, y, z)')

101
www.pdfgrip.com
M =

[2, 1, 0]
[1, 2, 0]
[0, 0, 2]

>> determ(M)

ans =

6

We see that the Hessian matrix is constant (it does not depend on the point (x,y,z )), therefore its value at the
origin is already found. The determinant is non-zero, so the origin is non-degenerate.

>> eigensys (M)

ans =

[1]
[2]
[3]

The Hessian matrix at the origin is positive definite, because all its eigenvalues are positive. Thus we conclude
that the origin is a minimum of the function.
4.9 Conditional Minima and Maxima. The Method of “Lagrange

Multipliers”
Suppose we want to optimize (i.e. maximize or minimize) the function f (x1,x2,...,xn), called the objective function, but
subject to certain restrictions given by the equations:
g1(x1,x2,...,xn ) = 0
g2(x1,x2,...,xn ) = 0
..........................
gk(x1,x2,...,xn ) = 0
This is the setting in which the Lagrangian is introduced. The Lagrangian is a linear combination of the objective
function and the constraints, and has the following form:
k
L(x1,x2,...,xn,l) = f (x1,x2,...,xn) + å l g (x , x
i =1
i i 1 2 ,..., x n )
The extreme points are found by solving the system by setting the components of the gradient vector of L to zero,
that is, Ñ L(x1,x2,...,xn,l) = (0,0,...,0). Which translates into:
æ¶L ¶L ¶L ¶L ¶L ¶Lö
ÑL =ç , ,..., , , ,..., ÷ = ( 0 , 0 ,......,0 )
¶
è 1 x ¶ x 2 ¶ x n ¶ l1 ¶ l2 ¶ ln ø
102
www.pdfgrip.com
By setting the partial derivatives to zero and solving the resulting system, we obtain the values of x1, x2,..., xn, l1,
l2,...,lk corresponding to possible maxima and minima.
To determine the nature of the points (x1, x2,..., xn) found above, the following bordered Hessian matrix is used:
é ¶2f ¶2f ¶2f ¶ gi ù

ê .......
ê ¶ x1
2
¶ x1 ¶ x 2 ¶ x1 ¶ x n ¶ x1 úú
ê ¶ f 2
¶2f ¶2f ¶ gi ú
ê ........ ú
ê ¶ x1 ¶ x 2 ¶ x 22 ¶ x2 ¶ xn ¶ x2 ú
ê ....... ....... ....... ....... ....... ú
ê ú
ê ¶2f ¶2f ¶2f ¶ gi ú
ê ....... ú
ê ¶ x1 ¶ x n ¶ x2 ¶ xn ¶ x n2 ¶ xn ú
ê ¶ gi ¶ gi ¶ gi ú
ê ....... 0 ú
ë ¶ x1 ¶ x2 ¶ xn û
The nature of extreme points can be determined by studying the set of bordered Hessian matrices:
é ¶2 f ¶2f ¶ gi ù
ê
é¶ f ¶ gi ù ê ¶ x1
2
¶ x1 ¶ x 2 ¶ x1 úú
ê¶ x2 ¶ x1 ú ê ¶2f ¶2 f ¶ g i ú ...
H1 = ê 1 ú H2 = ê ú Hn = H
ê ¶ gi ú ê ¶ x1 ¶ x 2 ¶ x 22 ¶ x2 ú
ê¶ x 0 ú
ë 1 û ê ¶g ¶ gi ú
ê i
0 ú
êë ¶ x1 ¶ x2 úû
For a single restriction g1, if H1 < 0, H2 < 0, H3 < 0,..., H < 0, then the extreme point is a minimum.
For a single restriction g1, if H1 > 0, H2 < 0, H3 > 0, H4 < 0, H5 > 0, ... then the extreme point is a maximum.
For a collection of restrictions gi(x1,..., xn) (i = 1, 2,..., k) the lower right 0 will be a block of zeros and the conditions
for a minimum will all have sign (-1)k, while the conditions for a maximum will have alternating signs with H1 having
sign (-1)k+1. When considering several restrictions at the same time, it is easier to determine the nature of the extreme
point by simple inspection.
MATLAB also offers some commands for function optimization, all of which require the prior use of the
command maple. Among them are the following:
maximize (expression) is the maximum value of the given expression with respect to
all of its variables
maximize(expression,{var1,...,varn}) is the maximum value of the given expression
with respect to the specified variables
minimize (expression) is the minimum value of the given expression with respect to all
of its variables
minimize(expression,{var1,...,varn}) is the minimum value of the given expression with
respect to the specified variables
minimize(expression,{var1,...,varn},{a1..b1,..., an...bn}) is the minimum value of the
given expression with respect to the specified variables, limiting the ranges of the
variables var1,..., varn to the intervals [a1, b1],..., [an, bn] (by default the variable
ranges are infinite)
103
www.pdfgrip.com
In addition, MATLAB has the command extrema, which gives us the extreme points of a function (the objective
function) subject to a few restrictions, determined by the method of Lagrange multipliers. Its syntax is as follows:

maple ('extrema(fobjective, {restrictions}, {variables},s)')

The variable s contains the list of candidate extreme points. To use this function one must read it from a Maple
library using the command maple (‘readlib(extrema)’). This library also includes the following commands (all
require the prior use of the command maple):
readlib (extrema) loads into the memory a library of commands for optimization with
restrictions
extrema(expression,equation,var) returns candidate extreme points for the function
defined by the given expression in the specified variable under the constraint defined
by the equation, using the method of Lagrange multipliers
extrema(expression,equation) finds candidate extreme points of the expression for all
variables subject to the restriction defined by the equation
extrema(expression, variable, {}) finds candidate extreme points for the expression in
the given variable and without restrictions
extrema(expression,{equation1,...,equationn},{variable1,...,variablen}) finds candidate
extreme points for the given expression with respect to the specified variables subject
to the constraints given by equation1,..., equationn
extrema(expression,{equ1,...,equn},{variable1,...,variablen},n) The variable n returns
the set of values obtained by evaluating the expression (objective function) at every
candidate extreme point. These values will be used to differentiate between minima
and maxima.
Exercise 4-20

f(x, y, z) = x + z subject to the constraint x 2 + y 2 + z 2 = 1.

>> maple ('readlib (extrema)');
>> maple ('extrema(x+z,{x^2+y^2+z^2=1},{x,y,z},s)')

ans =

{2 ^(1/2), - 2 ^(1/2)}

The command extrema returns the values that the objective function takes at each extreme point and the
variable s contains the extreme points.

>> maple ('s ')

ans =

{{z = 1/2 * 2 ^(1/2), x = 1/2 * 2 ^(1/2), y = 0}, {x = - 1/2 * 2 ^(1/2)}}
{{z = - 1/2 * 2 ^(1/2), y = 0}}

104
www.pdfgrip.com
Thus the extreme points are:

(Ö2/2, 0, Ö2/2) and (−Ö2/2, 0, −Ö2/2)
Now let’s determine the nature of these extreme points. To do this, we evaluate the objective function at these
points (we already know that the results are Ö2 and −Ö2).

>> maple ('obj:=(x,y,z) - > x+z');
>> maple('obj(sqrt(2)/2,0,sqrt(2)/2)')

ans =

2^(1/2)

>> maple('obj(-sqrt(2)/2,0,-sqrt(2)/2)')

ans =

-2 ^(1/2)

Then, at the point (−Ö2 /2, 0,−Ö2 /2) the function has a minimum, and there is a maximum at the point
(Ö2/2,0,Ö2/2).
Exercise 4-21
f (x , y ,z) = x 2 + y 2 - z
subject to the restrictions: x 2 + y 2 = 16 and x + y + z = 10.

>> maple('f:= (x^2+y^2) ^(1/2)-z');
>> maple('g1:= x^2+y^2=16; g2:= x+y+z = 10');
>> maple('readlib(extrema)');
>> maple('allvalues(extrema(f,{g1,g2},{x,y,z},s))')

ans =

{-6+4*2^(1/2)}, {-6-4*2^(1/2)}

>> maple('allvalues(s)')

ans =

{{y = 2 * 2 ^ (1/2) x = 2 * 2 ^ (1/2), z = - 4 * 2 ^(1/2) + 10}},
{{y = - 2 * 2 ^ (1/2), x = - 2 * 2 ^ (1/2), z = 4 * 2 ^(1/2) + 10}}

105
www.pdfgrip.com
Thus the extreme points are (Ö8, Ö8, −2Ö8 + 10) and (−Ö8, −Ö8, 2Ö8 + 10). Now let’s determine the nature of
these points. To this end, we evaluate the objective function at these points:

>> maple ('obj:=(x,y,z) - >(x^2+y^2) ^(1/2)-z');
>> numeric(maple('obj(-sqrt(8),-sqrt(8),sqrt(8)+10)'))

ans =

-8.8284

>> numeric(maple('obj(sqrt(8),sqrt(8),-sqrt(8)+10)'))

ans =

-3.1716

Thus, the function has a minimum at the point (−Ö8, −Ö8, 2Ö8 + 10) and a maximum at the point (Ö8, Ö8, − 2Ö8 + 10).
4.10 Some Applications of Maxima and Minima in Several

Variables
As in the case of a single variable, many practical problems involving several variables inevitably lead to the
consideration of derivatives.
Exercise 4-22
Find the dimensions of the rectangular cuboid of maximum volume which has a surface area of 10 square meters.
If x, y, z are the dimensions of the block, its volume will be V = xyz. As we know that it has a surface area of
10 square meters, the restriction will be 2xy + 2xz + 2yz = 10. We therefore have to maximize the objective
function V = xyz subject to the condition 2xy + 2xz + 2yz − 10 = 0. We will use the method of Lagrange multipliers:

>> maple('func:=x*y*z');
>> maple('g1:=2*x*y+2*x*z+2*y*z-10');
>> maple ('readlib (extrema)');
>> maple ('allvalues (extrema(func,{g1},{x,y,z}, a))')

ans =

{5/9 * 15 ^(1/2)}, {- 5/9 * 15 ^(1/2)}

>> maple ('a')

ans =

{{z = 1/3 * 15 ^(1/2), y = 1/3 * 15 ^(1/2), x = 1/3 * 15 ^(1/2)}},
{{z = - 1/3 * 15 ^(1/2), y = - 1/3 * 15 ^ (1/2) x = - 1/3 * 15 ^(1/2)}}

106
www.pdfgrip.com
Thus the extreme points are:

(−Ö15/3, −Ö15/3, −Ö15/3) and (Ö15/3, Ö15/3, Ö15/3).
The only point which is a meaningful solution is the second, because there can be no sides of negative length.
But it remains to be confirmed that (Ö15/3, Ö15/3, Ö15/3) is indeed a maximum:

>> maple ('obj:=(x,y,z) - > x * y * z');

>> numeric('[obj(-sqrt(15)/3,- sqrt(15)/3,- sqrt(15)/3), obj(sqrt(15)/3, sqrt(15)/3,
sqrt(15)/3)]')

ans =

-2.1517 2.1517

We see that the point is indeed a maximum, and so the maximal volume rectangular cuboid with fixed surface
area is a cube (x = y = z).
107
www.pdfgrip.com
Chapter 5
Vector Differential Calculus

and Theorems in Several Variables
5.1 Concepts of Vector Differential Calculus

Consider a function F : Rm → Rn:
(x1, x2,..., xm) → [F1(x1, x2,..., xm),..., Fn(x1, x2,..., xm)]

The vector function F is said to be differentiable at the point a =(a1,...,am) if each of the component functions
F1, F2,..., Fn is differentiable.
The Jacobian matrix of the above function is defined as:
é ¶ F1 ¶ F1 ¶ F1 ù
ê¶ x .......
¶ x2 ¶ xn ú
ê 1 ú
ê ¶ F2 ¶ F2 ¶ F2 ú
ê ....... ¶ ( F1 , F2 ,..., Fn )
J = ê ¶ x1 ¶ x2 ¶ xn úú =
¶ ( x1 , x 2 ,..., x n )
ê........ ....... ....... ....... ú
ê ú
ê ¶ Fn ¶ Fn
.......
¶ Fn ú
êë ¶ x1 ¶ x2 ¶ xn úû
The Jacobian of a vector function is an extension of the concept of a partial derivative for a single-component
function.
MATLAB has the command jacobian which enables you to calculate the Jacobian matrix of a function:
maple (‘jacobian([F1,F2,...,Fn],[x1,x2,...,xn])’) returns J = ∂ ( F1 , F2 ,..., Fn )
∂ ( x1 , x 2 ,..., x n )
109
www.pdfgrip.com
Chapter 5 ■ Vector Differential Calculus and Theorems in Several Variables
EXERCISE 5-1
Calculate the Jacobian matrix for the following function:

f ( x , y , z ) = [e x , cos( y ), sin(z )]
and find its value at the point (0,-p/2, 0).

>> sym (maple ('jacobian ([exp (x), cos (y), sin(z)], [x, y, z])'))
ans =

[exp (x), 0, 0]
[0,-sin (y), 0]
[0, 0, cos (z)]

>> maple ('m:=(x,y,z) - > array ([[exp(x), 0, 0], [0, -sin(y), 0], [0, 0, cos(z)]])');
>> sym(maple('m(0,-Pi/2,0)'))

ans =

[1, 0, 0]
[0, 1, 0]
[0, 0, 1]

We see that the Jacobian matrix evaluated at the given point is the identity matrix.
5.2 The Chain Rule

The chain rule allows you to differentiate the composition of vector functions. The chain rule is one of the most
familiar rules of differential calculus. It is often first introduced in the case of single variable real functions, and is then
generalized to vector functions. It says the following:
Suppose we have two vector functions
g : U Ì R n ® R m and f : V Ì R m ® R p .
where U and V are open and consider the composite function f  g : R n ® R p .
If g is differentiable at x 0 and f is differentiable at y 0 = g ( x 0 ) , then f  g is differentiable at x 0 and we have the
following:
D( f  g )( x 0 ) = Df ( y 0 ) Dg ( x 0 )
MATLAB will directly apply the chain rule when instructed to differentiate composite functions.
110
www.pdfgrip.com
EXERCISE 5-2
dg
Consider the functions f ( x , y ) = x 2 + y and h (u) = [sin(3u), cos(8u)] . If g (u) = f [h (u)] calculate at u = 0.
du
>> maple ('f:=(x,y) - > x ^ 2 + y ');
>> maple ('h: = u - > (sin(3*u), cos(8*u))');
>> maple ('g: = u - > (f@h) (u)');
>> pretty(sym(maple('diff(g(u),u)')))
6 sin(3 u) cos(3 u) - 8 sin(8 u)

This is the derivative of the composite function. Now, we find the value of its derivative at u = 0, defining it first as
a function of u:

>> maple ('dcomp: = u - > diff (g (u) u)');
dcomp: = u - > diff (g (u), u)

>> numeric (maple ('subs (u = 0, dcomp (u))'))

ans =

0
EXERCISE 5-3
¶z ¶z
Calculate and given that:
¶x ¶y
u2 + v2
z= , u = e x - y and v = e xy .
u2 - v2

>> maple ('f:=(u,v) - >(u^2+v^2) /(u^2-v^2)');

Now we define the following vector function:

>> maple ('g:=(x,y) - > (exp(x-y), exp(x*y))');

We define the function z as the composite of the previous two functions:

>> maple ('z:=(x,y) - > (f@g)(x,y)');

111
www.pdfgrip.com
Finally, we differentiate the composite function. MATLAB automatically differentiates composite functions
via the usual syntax:

>> pretty(sym(maple('simplify(diff(z(x,y),x))')))

(- 1 + y) exp(2 x - 2 y + 2 x y)
4 --------------------------------
2
(exp(2 x - 2 y) - exp(2 x y))

>> pretty(sym(maple('simplify(diff(z(x,y),y))')))

(1 + x) exp(2 x - 2 y + 2 x y)
4 -------------------------------
2
(exp(2 x - 2 y) - exp(2 x y))
Alternatively we could have directly calculated the composite function and differentiated it as follows:

>> pretty(sym(maple('simplify(f(g(x,y)))')))

exp(2 x - 2 y) + exp(2 x y)
---------------------------
exp(2 x - 2 y) - exp(2 x y)

Now, the problem would be reduced to finding the partial derivatives of this expression with respect to x and y,
which is trivial with MATLAB.

>> maple('comp:=(x,y)->simplify(f(g(x,y)))');
>> pretty(sym(maple('simplify(diff(comp(x,y),x))')))

(- 1 + y) exp(2 x - 2 y + 2 x y)
4 --------------------------------
2
(exp(2 x - 2 y) - exp(2 x y))

>> pretty(sym(maple('simplify(diff(comp(x,y),y))')))
(1 + x) exp(2 x - 2 y + 2 x y)
4 -------------------------------
2
(exp(2 x - 2 y) - exp(2 x y))

Obviously, the result obtained is the same.
112
www.pdfgrip.com
5.3 The Implicit Function Theorem

Consider the vector function F : A Ì Rn + m → Rm where A is an open subset of Rn + m
¾ [ F1 ( x , y ),..., Fm ( x , y )]
( x , y ) ¾®
F
If Fi (i = 1, 2,..., m) are differentiable with continuous derivatives up to order r and the Jacobian matrix
J = ∂ (F1,..., Fm) / ∂ (y1,..., ym) has non-zero determinant at a point ( x 0 , y 0 ) such that F ( x 0 , y 0 ) = 0 , then there is an
open U Ì Rn containing x 0 , an open V Ì Rm containing y 0 and a single-valued function f : U → V such that
F éë x , f ( x )ùû = 0 "x Î U and f is differentiable of order r with continuous derivatives.
This theorem guarantees the existence of certain derivatives of implicit functions. MATLAB allows differentiation
of implicit functions and offers the results in those cases where the hypothesis of the theorem are met.
EXERCISE 5-4
Find the conditions on x, y, z so that the surface defined by the following equation defines an implicit function
z = z(x,y) where z is a differentiable function:
x 3 + 3 y 2 + 8 xz 2 - 3 yz 3 = 1
Calculate:
¶z ¶z ¶ 2 z ¶2z
, , 2 and 2 .
¶x ¶y ¶x ¶y

>> maple('f:=(x,y,z)->x^3+3*y^2+8*x*z^2-3*z^3*y-1');

In order for the hypotheses of the implicit function theorem to be met, the partial derivative of f with respect to the
variable z must be non-zero.

>> pretty(sym(maple('diff(f(x,y,z),z)')))
2
16 x z - 9 z y

Thus the required condition is 16xz ¹ 9yz 2.
Now we can calculate the partial derivatives, assuming that the above condition is met.
The implicit (total) derivative of f and D(z) are found as follows:

>> pretty(sym(maple('D(f(x,y,z)=0)')))

2 2 2 3
3 D(x) x + 6 D(y) y + 8 D(x) z + 16 x D(z) z - 9 D(z) z y - 3 z D(y) = 0

>> pretty(sym(maple('D(z):=solve(D(f(x,y,z)=0),D(z))')))
>> pretty(sym(maple('D(z)')))
2 2 3
3 D(x) x + 6 D(y) y + 8 D(x) z - 3 z D(y)
D(z) := - ---------------------------------------------
2
16 x z - 9 z y

113
www.pdfgrip.com
As y does not depend on x (x and y are, by hypothesis, independent variables, and the variable z depends on
x and y ), we put D(y) = 0 in the expression for D(z)/D(x):

>> pretty(sym(maple('simplify(subs(D(y)=0,D(z)/D(x)))')))
2 2
3 x + 8 z
- -------------------
z (16 x - 9 z y)

This gives the expression for ∂z/∂x. For ∂z/∂y, we similarly calculate the following:

>> pretty(sym(maple('simplify(subs(D(x)=0,D(z)/D(y)))')))
3
2 y - z
- 3 ------------------
z (16 x - 9 z y)

To calculate ∂ 2z /∂x 2, we find the partial derivative of ∂z/∂x with respect to x:

>> pretty(simple(sym(diff(maple('simplify(subs(D(y)=0,D(z)/D(x)))'),'x'))))
2 2
24 x - 27 x z y - 64 z
- 2 ----------------------------
2
z (16 x - 9 z y)

To calculate ∂ 2z /∂y 2, we find the partial derivative of ∂z/∂y with respect to y:

>> pretty(simple(sym(diff(maple('simplify(subs(D(x)=0,D(z)/D(y)))'),'y'))))
4
32 x - 9 z
- 3 -------------------
2
z (16 x - 9 z y)
EXERCISE 5-5
Show that near the point (x, y, u, v ) = (1,1,1,1) the system
xy + yvu 2 = 2
xu 3 + y 2v 4 = 2
can be solved uniquely for u and v as functions of x and y (u = u(x,y), v = v (x,y)).

First, we check that the hypotheses of the implicit function theorem are met at the point (1,1,1,1).
114
www.pdfgrip.com
The functions are differentiable and have continuous derivatives. We need to show that the corresponding
Jacobian determinant is non-zero at the point (1,1,1,1):

>> maple('f1:=(x,y,u,v)-> x*y +y*v*u^2-2');
>> maple('f2:=(x,y,u,v)->x*u^3+y^2*v^4-2');
>> sym(maple('jacobian([f1(x,y,u,v),f2(x,y,u,v)],[u,v])'))

ans =

[2*y*v*u, y*u^2]
[3*x*u^2, 4*y^2*v^3]

>> J=sym(maple('subs(x=1,y=1,u=1,v=1,jacobian([f1(x,y,u,v),f2(x,y,u,v)],[u,v]))'))

J =

[2, 1]
[3, 4]

>> determ(J)
ans =

5

Thus the assumptions of the implicit function theorem are met and the proposed system can be solved uniquely.
5.4 The Inverse Function Theorem

Consider the vector function f : U Ì R n → R n where U is an open subset of R n
(x1, x2,..., xn) → [f1(x1, x2,..., xn),..., fn(x1, x2,..., xn)]
and assume it is differentiable with continuous derivative.
If there is an x 0 such that |J | = |∂(f1,...,fn) / ∂(x1,...,xn)| ¹ 0 at x0, then there is an open set A containing x 0 and an
open set B containing f ( x 0 ) such that f ( A) = B and f has an inverse function f -1 : B ® A that is differentiable with
continuous derivative. In addition we have:
-1
Df -1 ( y ) = éë Df ( x )ùû and if J = ∂ (f1,..., fn) / ∂ (x1,..., xn) then |J -1| = 1 / |J |.
MATLAB automatically performs the calculations related to the inverse function theorem, provided that the
assumptions are met.
115
www.pdfgrip.com
EXERCISE 5-6
Given the vector function (u(x,y), v(x,y)) where:

x4 + y4
u( x , y ) = , v( x , y ) = sin( x ) + cos( y )
x
find the conditions under which the inverse vector function (x(u,v ), y(u,v )) with x = x(u, v ) and y = y(u,v ) exists
and find the derivative of the inverse transformation. Find its value at (p/4,-p/4).
The conditions to be met are precisely the hypotheses of the inverse function theorem. The functions are
differentiable with continuous derivatives, except perhaps at x= 0. Now let’s consider the Jacobian of the
transformation ∂(u(x,y), v(x,y)) / ∂(x, y):

>> J=simple(sym(maple('jacobian([(x^4+y^4)/x,sin(x)+cos(y)],[x,y])')))

J =

[(3*x^4-y^4)/x^2, 4*y^3/x]
[ cos(x), -sin(y)]

>> pretty(sym(determ(J)))

4 4 3
3 sin(y) x - sin(y) y + 4 y cos(x) x
- ---------------------------------------
2
x

Therefore, at the points where this expression is non-zero, x and y can be expressed in terms of u and v.
In addition, we must have x ¹ 0.
We calculate the derivative of the inverse function. Its value is the inverse of the Jacobian matrix and the
determinant of the inverse Jacobian matrix is the reciprocal of the determinant of the Jacobian matrix:

>> I=simple(inv(J))

I =

[sin(y)/(3*sin(y)*x^4-sin(y)*y^4+4*y^3*cos(x)*x)*x^2,
4*y^3*x/(3*sin(y)*x^4-sin(y)*y^4+4*y^3*cos(x)*x)]
[cos(x)/(3*sin(y)*x^4-sin(y)*y^4+4*y^3*cos(x)*x)*x^2,
-(3*x^4-y^4)/(3*sin(y)*x^4-sin(y)*y^4+4*y^3*cos(x)*x)]

>> pretty(sym(simple(determ(inv(J)))))

2
x
-------------------------------------------
4 4 3
3 sin(y) x - sin(y) y + 4 y cos(x) x

116
www.pdfgrip.com
Observe that this is indeed the reciprocal of the determinant found above.
Next we find the value of the inverse Jacobian at the point (p/4,−p/4) :

>> numeric(subs(subs(determ(I),pi/4,'x'),-pi/4,'y'))

ans =

0.3821

>> numeric(subs(subs(symdiv(1,determ(J)),pi/4,'x'),-pi/4,'y'))

ans =

0.3821

Here it is clear that the determinant of the Jacobian of the inverse function is the reciprocal of the determinant of
the Jacobian of the function.
EXERCISE 5-7
Demonstrate that the transformation between cartesian and polar coordinates complies with the assumptions of
the inverse function theorem.
We know that the transformation equations are:
x = a cos[b], y = a sin[b]
Obviously, the functions have continuous partial derivatives. Let's see if the determinant of the Jacobian of the
transformation is non-zero:

>> J = simple (sym (maple ('jacobian ([a * cos (b), a * sin (b)], [a, b])')))

J =

[cos (b),-a * sin (b)]
[sin(b), a*cos(b)]
>> pretty(simple(sym(determ(J))))
a

We see that the Jacobian of the transformation is non-zero (a ¹ 0). Thus, the inverse function theorem is
applicable. The determinant of the Jacobian of the inverse transformation will be 1/a.
117
www.pdfgrip.com
5.5 The Change of Variables Theorem

Suppose we have a function f(x,y) that depends on the variables x and y, and that meets all the conditions of
differentiation and continuity necessary for the inverse function theorem to hold. We introduce new variables u and v,
relating to the above, regarding them as functions u = u(x,y) and v = v(x,y), so that u and v also fulfil the necessary
conditions of differentiation and continuity (described by the inverse function theorem) to be able to express x and y
as functions of u and v: x=x (u,v) and y=y(u,v).
Under the above conditions, it is possible to express the initial function f as a function of the new variables u and
v using the expression:
f(u,v) = f (x(u,v), y(u,v))|J| where J is the Jacobian ∂ (x(u, v), y(u,v))/∂(u, v).
The theorem generalizes to vector functions of n components.
EXERCISE 5-8
Consider the function f ( x , y ) = e x - y and the transformation u = u(x,y ) = x + y, v = v(x,y ) = x. Calculate f (u,v ).

>> maple ('f:=(x,y) - > exp(x-y)');

This transformation fulfils the conditions of the inverse function theorem (continuous partial derivatives and a
Jacobian with non-zero determinant).
We find the inverse transformation and its Jacobian to apply the change of variables theorem:

>> maple ('solve({u=x+y,v=x},{x,y})')

ans =

{x = v, y = u-v}

>> sym (maple ('jacobian([v,u-v],[u,v])'))

ans =

[0, 1]
[1, - 1]

>> pretty (simple (sym (maple ('f (v, u-v) * abs (det (jacobian([v,u-v],[u,v])))'))))
exp(2 v-u)

The requested function is f(u,v)= e2v-u.
118
www.pdfgrip.com
5.6 Taylor’s Theorem with n Variables

Let f : R n ® R , (x1,...,xn) → f (x1,...,xn), be differentiable k times with continuous partial derivatives.
The Taylor series expansion of order k of f ( x ) at the point a = (a1 ,..., an ) is as follows:
n
¶f 1 n n ¶2f
f ( x ) = f (a ) + å (a )t i + å å (a )t it j +
i =1 ¶ x i 2 ! i =1 j =1 ¶ xi¶ x j
1 n n n
¶3f
åå
3!! i =1 j =1
å
k =1 ¶ xi¶ x j ¶ x k
(a )tit jt k + ... + R(k + 1)
Here
x = ( x1 , x 2  x n ), a = (a1 , a2 an ),ti = xi - ai (i = 1, 2 ,n)
R = remainder.
Normally, the series are given up to order 2.
EXERCISE 5-9
Find the Taylor series of order 2 at the point (1,0) of the function:
2
f ( x , y ) = e ( x -1) cos( y ).

>> maple('f:=(x,y)->exp((x-1)^2)*cos(y)');

>> pretty(sym(maple('f(1,0)+D[1](f)(1,0)*(x-1)+D[2](f)(1,0)*(y)+
+(1/2!)*(D[1,1](f)(1,0)*(x-1)^2+ 2*D[1,2](f)(1,0)*( x-1)*y+D[2,2](f)(1,0)*(y)^2)')))

2 2
1 + (x - 1) - 1/2 y
EXERCISE 5-10
Find the Taylor series of order 2 at the origin of the function:

2
f (x , y) = e x+y

>> maple('f:=(x,y)->exp(x+y^2)');
>> pretty(sym(maple('f(0,0)+D[1](f)(0,0)*(x)+D[2](f)(0,0)*(y)+
(1/2!)*(D[1,1](f)(0,0)*(x)^2+2*D[1,2](f)(0,0)*(x)*( y)+D[2,2](f)(0,0)*(y)^2)')))

2 2
1 + x + 1/2 x + y
119
www.pdfgrip.com
5.7 Vector Fields. Curl, Divergence and the Laplacian

We recall some basic definitions:
Definition of gradient: If h = f(x,y,z), then the gradient of f, which is denoted by Df ( x , y , z ) , is the vector:
¶f ( x , y , z ) ¶f ( x , y , z ) ¶f ( x , y , z )
grad( f ) = Df ( x , y , z ) = i+ j+ k
¶x ¶y ¶z
Definition of a scalar potential of a vector field: A vector field F is called conservative if there is a differentiable
function f such that F = Df . The function f is known as a scalar potential function for F .
Definition of the curl of a vector field: The curl of a vector field F ( x , y , z ) = Mi + Nj + Pk is the following:
æ ¶P ¶N ö æ ¶P ¶M ö æ ¶N ¶M ö
curl F ( x , y , z ) = D ´ F ( x , y , z ) = ç - ÷i - ç - ÷ j +ç - ÷k
è ¶y ¶z ø è ¶x ¶z ø è ¶x ¶y ø
Definition of a vector potential of a vector field: A vector field F is a vector potential of another vector field
G if F = curl (G).
Definition of the divergence of a vector field: The divergence of the vector field F ( x , y , z ) = Mi + Nj + Pk is the
following:
¶M ¶N ¶P
diverge F ( x , y , z ) = D × F ( x , y , z ) = + +
¶x ¶y ¶z
Definition of the Laplacian: The Laplacian is the differential operator defined by:
¶2 ¶2 ¶2
Laplacian = D 2 = D × D = + +
¶x 2 ¶y 2 ¶z 2
MATLAB provides the following commands for the above operations:

maple (‘grad(f,[x1,...,xn])’) calculates the gradient of f(x1,...,xn)
maple(‘potential([f1,...,fn],[x1,...,xn],p)’) determines if the vector field with
components [f1,..., fn] has a scalar potential (to do this, the vector field has to be
irrotational or have zero curl). If so, p is the scalar potential.
maple (‘curl([f1,...,fn],[x1,...,xn])’) gives the rotational component of the vector field
[f1,..., fn].
maple(‘vectpotent([f1,...,fn],[x1,...,xn],p)’) determines if the vector field with
components [f1,..., fn] has a vector potential (for this purpose, the vector field has to be
incompressible or have zero divergence). If so, p is the vector potential.
maple (‘diverge([f1,...,fn],[x1,...,xn])’) gives the divergence of the vector field with
components [f1,..., fn].
maple (‘laplacian(f,[x1,...,xn])’) computes the Laplacian of the function f.
120
www.pdfgrip.com
EXERCISE 5-11
Calculate the gradient and Laplacian of the function

1
w= .
2
1- x 2 - y - z 2

>> maple('w:=1/sqrt(1-x^2-y^2-z^2)');
>> pretty(simple(sym(maple('grad(w,[x,y,z])'))))
x y z
[---------------------, ----------------------, ---------------------]
2 2 2 3/2 2 2 2 3/2 2 2 2 3/2
(1 - x - y - z ) (1 - x - y - z ) (1 - x - y - z )
>> pretty(simple(sym(maple('laplacian(w,[x,y,z])'))))

3
---------------------
2 2 2 5/2
(1 - x - y - z )
EXERCISE 5-12
Calculate the scalar potential of the vector field [1,2 y ,3z 2] .

>> maple ('potential([1,2*y,3*z^2],[x,y,z],P)')
ans =

true

The response true is given because there is a scalar potential, i.e., because the curl of the given vector
field is null:

>> sym (maple ('curl([1,2*y,3*z^2],[x,y,z])'))
ans =

[0, 0, 0]

The value of the scalar potential is contained in the variable P :

>> pretty(sym(maple('P')))

2 3
y + x + z

121
www.pdfgrip.com
If we calculate the gradient of this scalar potential, we get the initial vector field.

>> pretty(simple(maple('grad(y^2+x+z^3,[x,y,z])')))

2
[1, 2 y, 3 z ]
EXERCISE 5-13
Calculate the curl and the divergence of the vector field:

x
F ( x , y , z ) = tan -1 i + ln x 2 + y 2 j + k .
y

>> pretty(simple(sym(maple('curl([arctan(x/y),ln(sqrt(x^2+y^2)),1],[x,y,z])'))))

x
[0, 0, 2 -----------]
2 2
x + y

>> pretty(simple(sym(maple('diverge([arctan(x/y),ln(sqrt(x^2+y^2)),1],[x,y,z])'))))

y
2 ---------
2 2
x + y
EXERCISE 5-14
Calculate the vector potential of the following vector fields:
F = [ xz , - yz , y ], G = [ xe y , x cos(z ), - ze y ], H = [ x cos( y ), - sin( y ),sin( x )].

>> maple ('vecpotent([x*z,-y*z,y],[x,y,z],Q)')
ans =

true

Thus, we know that there is a vector potential, and the divergence of the vector field is zero:

>> sym (maple ('diverge([x*z,-y*z,y],[x,y,z])'))
ans =

0

122
www.pdfgrip.com
The value of the vector potential is contained in the variable Q:

>> pretty(simple(sym(maple('print(Q)'))))
2 2
[- 1/2 y (z + y), - 1/2 x z , 0]

If we calculate the curl of the previous result, we obtain the initial vector field:

>> pretty (sym (maple ('curl ([-1/2 * y *(z^2+y), - 1/2 * x * z ^ 2, 0], [x, y, z])')))
[x z,- y z, y]

Now, we calculate the vector potential of the other two fields:

>> pretty(simple(sym(maple('print(Q)'))))
2 2
[- 1/2 y (z + y), - 1/2 x z , 0]
[-z sin (y) - y sin (x), - x z cos(y), 0]

If we calculate the curl of these vector potentials, we obtain the initial vector fields:

>> pretty(sym(maple('curl([- sin(z)*x, - x*exp(y)*z, 0],[x,y,z])')))
[x exp(y), - x cos(z), - z exp(y)]
>> pretty(sym(maple('curl([- sin(y)*z - sin(x)*y, - x*cos(y)*z, 0],[x,y,z])')))
[x cos(y), - sin(y), sin(x)]
5.8 Coordinate Transformation

In this section we will see how to convert cylindrical and spherical coordinates to rectangular coordinates, cylindrical
coordinates to spherical coordinates, and their inverse transformations. Given the simplicity of the functions we will
define them ourselves.
First we recall the definitions of cylindrical and spherical coordinates:
In a cylindrical coordinate system, a point P in the space is represented by a triplet (r, q, z), where:
r is the distance from the origin (O) to the projection P' of P in the XY plane
q is the angle between the X axis and the segment OP'
z is the distance PP'
In a spherical coordinate system, a point P in the space is represented by a triplet (r, q, f), where:
r is the distance from P to the origin
q is the same angle as the one used in cylindrical coordinates
f is the angle between the positive Z axis and the segment OP
123
www.pdfgrip.com
The following conversion equations are easily found:

Cylindrical to rectangular:
x = r cosq
y = r sinq
z=z
Rectangular to cylindrical:
r = x2 + y2
y
q = tan -1
x
z=z
Spherical to rectangular:
x = r sin f cosq
y = r sin f sin q
z = r cosf
Rectangular to spherical:
r = x2 + y2 + z2
y
q = tan -1
x
z
f = cos -1
x + y2 + z2
2
Therefore, in MATLAB we can define the following functions:

>> maple ('sphrec:=(r,t,f) - > [r * sin (f) * cos (t), r * sin (f) * sin (t), r * cos (f)]');

>> maple ('cylrec:=(r,t,z) - > [r * cos (t), r *sin (t), z]');
>> maple ('recsph:=(x,y,z) - > (sqrt(x^2+y^2+z^2), arctan (y/x), arccos (z/sqrt(x^2+y^2+z^2)))');

>> maple ('reccyl:=(x,y,z) - > [sqrt(x^2+y^2), arctan (y/x), z]');
EXERCISE 5-15
Express the point given in spherical coordinates by (4,p/6,p/4) in rectangular and cylindrical coordinates.

>> maple ('sphrec(4,pi/6,pi/4)')

ans =

[2 ^(1/2) * 3 ^(1/2), 2 ^(1/2), 2 * 2 ^(1/2)]

124
www.pdfgrip.com
Now, we convert the rectangular to cylindrical coordinates:

>> maple ('reccyl (2 ^(1/2) * 3 ^(1/2), 2 ^(1/2), 2 * 2 ^(1/2))')

ans =

[2 * 2 ^ (1/2), 1/6 * pi, 2 * 2 ^(1/2)]
EXERCISE 5-16
Express the point given in cylindrical coordinates by (4,5p/6,3) in rectangular and spherical coordinates.

>> maple('cylrec(4,5*Pi/6,3)')

ans =

[- 2 * 3 ^ (1/2), 2, 3]

Now, we convert the rectangular to spherical coordinates:

>> maple ('recsph (- 2 * 3 ^ (1/2), 2, 3)')
ans =

5, - 1/6 * pi, acos(3/5)
EXERCISE 5-17
Express the equations of the surfaces xz = 1 and x 2 + y 2 + z 2 = 1 in spherical coordinates.

>> pretty(sym(maple('subs(x=r*sin(a)*cos(t),y=r*sin(a)*sin(t),z=r*cos(a), x*z=1)')))

2
r sin(a) cos(t) cos(a) = 1

>> pretty(simple(sym(maple('subs(x=r*sin(a)*cos(t),y=r*sin(a)*sin(t),
z=r*cos(a), x^2+y^2-z^2=1)'))))
2
- r cos(2 a) = 1
125
www.pdfgrip.com
EXERCISE 5-18
Express in cartesian and spherical coordinates the equation of the surface which in cylindrical coordinates is
given by z = r 2 (1 + sin (t )).

>> pretty(sym(maple('expand(simplify(subs(r=sqrt(x^2+y^2),t=arctan(y/x),
z=z, z=r^2*(1+sin(t))),symbolic))')))
2 2 2 2 1/2
z = x + y + (x + y ) y

Now let us convert the cartesian to spherical coordinates:

>> pretty(sym(maple('assume(t,real):expand(simplify(subs(x=r*sin(a)*cos(t),
y=r*sin(a)*sin(t), z=r*cos(a), z=(x^2+y^2)*(1+sin(atan(y/x)))),symbolic))')))
2 2 2 2 2 2
r cos(a) = r sin(t) - r cos(a) sin(t) + r - r cos(a)
EXERCISE 5-19
Find the unit tangent, the unit normal, and the unit binormal vectors of the twisted cubic: x = t, y = t 2, z = t 3.
We begin by restricting the variable t to the real field:

>> maple('assume(t,real)') ;

We define the symbolic vector V as follows:

>> syms t, V=[t,t^2,t^3]

V =

[t, t ^ 2, t ^ 3]

The tangent vector is calculated by:

>> tang= diff (V)

tang =

[1, 2 *, 3 * t ^ 2]

126
www.pdfgrip.com
The unit tangent vector will be:

>> ut = simple (tang / sqrt (dot(tan,tan)))

ut =

[1/(1+4*t^2+9*t^4)^(1/2),2*t/(1+4*t^2+9*t^4)^(1/2),3*t^2/(1+4*t^2+9*t^4)^(1/2)]

To find the unit normal vector we calculuate ((v'∧v'') ∧v') /(|v'∧v''| |v'|):

>> v1=cross(diff(V),diff(V,2)) ;
>> un=simple(cross(v1,tang)/(sqrt(dot(v1,v1))*sqrt(dot(tang,tang))))
un =

[ (-2*t-9*t^3)/(9*t^4+9*t^2+1)^(1/2)/(1+4*t^2+9*t^4)^(1/2),
(1-9*t^4)/(9*t^4+9*t^2+1)^(1/2)/(1+4*t^2+9*t^4)^(1/2),
(6*t^3+3*t)/(9*t^4+9*t^2+1)^(1/2)/(1+4*t^2+9*t^4)^(1/2)]

The unit binormal vector is the vector product of the tangent vector and the unit normal vector.

>> ub=simple(cross(ut,un))
ub =

[3*t^2/(9*t^4+9*t^2+1)^(1/2),-3*t/(9*t^4+9*t^2+1)^(1/2),1/(9*t^4+9*t^2+1)^(1/2)]

The unit binormal vector can also be calculated via (v'∧v ") / |v'∧v" | as follows:

>> ub=simple(v1/sqrt(dot(v1,v1)))
ub =

[3*t^2/(9*t^4+9*t^2+1)^(1/2),-3*t/(9*t^4+9*t^2+1)^(1/2),1/(9*t^4+9*t^2+1)^(1/2)]

We have calculated the Frenet frame for a twisted cubic.
127
www.pdfgrip.com
Chapter 6
Integration and Applications
MATLAB works with integral calculus in a clear and simple way. The number of functions that enables you to work in
this area is not very high, but they are very efficient in solving integration problems. You can calculate the indefinite
integral of most integrable functions whose structure is not very complicated; for example, functions that involve
simple logarithms, exponentials, rational, trigonometric functions, inverse trigonometric functions, etc. Definite and
improper integrals do not present problems for MATLAB. Double integrals, triple integrals and n-fold integrals are
also easily found.
The most commonly used MATLAB commands for integral calculus are:
syms x, int(f(x), x) or int(‘f(x)’, ‘x’)
Computes the indefinite integral ò f ( x )dx
int (int (‘f(x,y)’, ‘x’), ‘y’)
Computes the double integral ò ò f ( x , y )dxdy
syms x y int (int (f(x,y), x), y)
Computes the double integral òò f ( x , y )dxdy
int(int(int(...int(‘f(x,y...z)’, ‘x’), ‘y’)...), ‘z’)
Computes ò ò ¼ ò f ( x , y ,¼, z ) dxdy ¼dz
syms x y z, int(int(int(...int(f(x,y...z), x), y)...), z)
syms x a b, int(f(x), x, a, b)
b
Computes the definite integral òa
f ( x )dx
int (‘f (x)’, ‘x’, ‘a’, ‘b’)
b
f ( x )dx
int(int(‘f(x,y)’, ‘x’, ‘a’, ‘b’), ‘y’, ‘c’, ‘d’))
b d
Computes the definite double integral ò ò
a c
f ( x , y )dxdy
syms x y b c d, int (int (f(x,y), x, a, b), y, c, d)
b d
Computes the definite double integral ò ò
a c
f ( x , y )dxdy
int(int(int(....int(‘f(x,y,...,z)’, ‘x’, ‘a’, ‘b’), ‘y’, ‘c’, ‘d’),...), ‘z’, ‘e’, ‘f’)
b d f
Computes ò ò ¼ ò f ( x , y ,¼ , z ) dxdy ¼ dz
a c e
129
www.pdfgrip.com
Chapter 6 ■ Integration and Applications
Syms x y z a b c d e f, int (int (int (... int (f(x,y,...,z), x, a, b), y, c, d),...), z, e, f)

b d f
a c e
In addition, with the prior use of the command maple, the following commands can be used for integration:
int (f (x), x) Computes the indefinite integral ò f ( x )dx
b
int (f (x), x = a...b) Computes the definite integral òa
f ( x )dx
int(f(x), x = a..b, continuous)
Computes the definite integral removing the check for continuity of the
function in the interval (a, b)
residue (f(x), x = a)
Returns the algebraic residue of f(x) at the point a. Used to solve
integrals by the method of residues. The algebraic residue is the
coefficient of x -1 in the Laurent expansion of f(x) at a.
readlib (singular):singular(f(x))
Returns the singularities of the function f(x)
readlib (singular): singular (f (x), x)
Returns the singularities of the function f(x) treating all variables other
than x as constants
readlib (singular): singular (f(x, y, z,...), {x, y,z,...})
Returns the singularities of the function f(x,y,z,...) treating all variables
other than x, y,z,... as constants
op (Int (f (x), x))
Shows the internal structure of the Int command and allows access to
its operands
op (n, Int (f (x), x)
Removes the n-th operand of int. The first operand is the integrand, the
second operand is the variable of integration or x = a...b for definite
integrals, and the third operand is continuous, if it is being used. The
operands can be replaced with the command subsop.
int (int (f(x,y), x), y) computes the indefinite integral ò ò f ( x , y )dxdy
int(int(int(...int(f(x,y,...,z), x), y),...), z)
int(int(f(x,y),x=a..b),y=c..d)
b d
Computes the definite integral ò ò
a c
f ( x , y )dxdy
int (int (int (... int (f(x,y,...,z), x = a...b), y = c...d),...), z = e...f)
b d f
a c e
130
www.pdfgrip.com

>> int('1/(x^2-1)','x')

ans =

−atanh(x)

>> pretty(simple(int('a*log(1+b*x)','x')))

a (log(1 + b x) - 1) (1 + b x)
------------------------------
B

>> syms x a b, pretty(simple(int(a*log(1+b*x),x)))

a (log(1 + b x) - 1) (1 + b x)
------------------------------
B

>> pretty(int('x^n','x'))
(n + 1)
x
---------
n + 1

We have seen how MATLAB allows you to include parameters in the integrals, which are treated as generic
constants:

>> pretty(simple(int(int('x^3+y^3','x',0,'a'),'y',0,'b')))

4 4
1/4 a b + 1/4 a b

>> syms x y a b, pretty(simple(int(int(x^3+y^3,x,0,a),y,0,b)))

4 4
1/4 a b + 1/4 a b

>> pretty(simple(int(int('x^3+y^3','x',0,'a'),'y',0,'x')))

4 4
1/4 a x + 1/4 a x

In these last examples definite integrals have been solved, one of which has a variable limit of integration,
which is perfectly valid with MATLAB.
131
www.pdfgrip.com
6.1 Indefinite Integrals

In the examples that follow, we demonstrate the ease with which MATLAB can find integrals that are not so obvious
using classical techniques. MATLAB is able to find indefinite integrals involving simple, rational, exponential,
logarithmic, trigonometric and inverse trigonometric functions.
Exercise 6-1
Find the following integrals:

ò sec( x )csc ( x ) dx , òx cos ( x ) dx , ò cos ( 2 x ) dx.
-1

>> pretty(simple(int('sec(x)*csc(x)')))

log(tan(x))
>> pretty(simple(int('x*cos(x)')))

cos(x) + x sin(x)

>> pretty(simple(int('acos(2*x)')))
2 1/2
x acos(2 x) - 1/2 (1 - 4 x )
Exercise 6-2
Find the following integrals:

1 ln( x ) 1
ò x2 +1
dx , ò x ò ( 2 + x) 1 + x
, dx , òx 3
1 + x 4 dx

>> pretty(simple(int('1/sqrt(x^2+1)')))

asinh(x)

>> pretty(simple(int('log(x)/x^(1/2)')))

1/2 1/2
x (4 log(x ) - 4)

>> pretty(simple(int('1/((2+x)*(1+x)^(1/2))')))

1/2
2 atan((x + 1) )

>> pretty(simple(int('x^3*sqrt(1+x^4)')))

4 3/2
1/6 (1 + x )
132
www.pdfgrip.com
6.2 Integration by Substitution (or Change of Variables)

Integration by substitution is used to solve integrals ò f ( x ) dx where a direct solution is not immediate and where
other techniques fail.
The method is to find a function x = g (t) which replaces x under the integral sign, converting the integral into
another simpler integral in the variable t. The new integrand will in the majority of cases be immediately integrable or
will be a rational function.
The substitution x = g(t) must meet the following conditions:
1. g(t) is differentiable:
dx
= g ¢ ( t ).
dt
2. x = g(t) must be invertible, supporting an inverse function t = h (x).
Once the transformed integral has been calculated as a function of t, we must substitute t = h (x) to give the final
solution as a function of x.
The substitution method is one of the most wide ranging of integral calculus, because of the great variety of
substitutions that can be used, depending on the type of function that appears under the integral sign. However, it can
easily go wrong in the sense that the transformed integral may be more difficult than the original integral, so some
care is needed to choose an appropriate substitution.
MATLAB provides the function maple(‘changevar’) in the student library (which has to be previously loaded via
the command with (student)). This allows you to solve integrals by substitution. Its syntax is as follows:
maple (‘changevar(expr_vnew=expr_vold, integral, v_new))
This performs the change of variable (as specified by the expression
in the new variable equal to the expression in the old variable) for the
given integral using the new variable
As an example, we calculate the integral of the function 1 / (1 + e x) via the change of variable t = e x:

>> maple('with(student)');
>> pretty(simple(sym(maple('changevar(t=exp(x),int(1/(1+exp(x)),x),t)'))))

-log (1 + exp (x)) + x

Exercise 6-3
Solve the following integrals by substitution:
æxö
tan -1 ç ÷
è2ø a sin( x ) cos3 ( x )
ò x 2 + 4 dx , ò (1 - x 2 )3/2 dx , ò sin( x ) dx.

>> pretty(simple(sym(maple('changevar(t=atan(x/2),int(atan(x/2)/(4+x^2),x),t)'))))

2 2
- 1/16 log(1 + 1/2 i x) + 1/8 log(2) log(1/16 x + 1/4)

2
- 1/8 dilog(1/2 + 1/4 i x) - 1/16 log(1 - 1/2 i x)

- 1/8 dilog(1/2 - 1/4 i x)

133
www.pdfgrip.com
Results can easily be found by using the integral operator in inert mode, i.e. Int, making the substitution and
finding the value of the integral later with value as follows:

>> pretty(sym(maple('value(simplify(changevar(t=arctan(x/2),
Int(arctan(x/2)/(4+x^2),x),t)))')
2
1/4 t

Now we undo the change of variable to obtain the final result:

>> pretty(subs('1/4*t^2','atan(x/2)'))
2
1/4 atan(1/2 x)

>> pretty(simple(sym(maple('changevar(t=asin(x), Int(asin(x)/(1-x^2)^(3/2),x),t)'))))

t tan(t) + log(cos(t))

Now we reverse the change of variable:

>> pretty(subs('t*tan(t)+t*log(t)','asin(x)'))

asin(x) x
----------- + asin(x) log(asin(x))
2 1/2
(1 - x )

>> pretty(simple(sym(maple('changevar(t=sin(x), Int(cos(x)^3/sin(x)^(1/2),x),t)'))))

1/2 2
- 2/5 t
- 5) (t

>> pretty(subs('-2/5*t^(1/2)*(t^2-5)','sin(x)'))

1/2 2
- 2/5 sin(x) (sin(x) - 5)
134
www.pdfgrip.com
Exercise 6-4
Solve the following integrals:

ó dx ó 9 - 4x 2 1
, dx , ò x ( 3 + 5 x ) 4 dx
ô ô 8 3
ô ô
õ x 3 x 2 + 3x - 1 õ x
For the first integral we use the change of variable 1/x = t.

>> pretty(simple(sym(maple('changevar(t=1/x, int(x^(-3)*(x^2+3*x-1)^(-1/2),x),t)'))))

2 1/2 2 1/2 1/2
(x + 3 x - 1) (x + 3 x - 1) (3 x - 2) 13
1/2 ----------------- + 9/4 ----------------- + 31/8 asin(1/13 ---------------)
2 x x
x

For the second integral, we use the change of variable x = 3/2sin(t ), or equivalently, t = arcsin(2x/3):

>> pretty(simple(sym(maple('changevar(t=asin(2*x/3), int(x^(-1)* (9-4*x^2)^(-1/2), x),t)'))))
3
- 1/3 atanh(-------------)
2 1/2
(9 - 4 x )

For the third integral, we use the change of variable 3 + 5x 3 = t 4:

>> pretty(simple(sym(maple('changevar(t=(3+5*x^3)^(1/4),int(x^8* (3+5*x^3)^(1/4),x),t)'))))
3 6 9 3 1/4
4/73125 (288 - 120 x + 125 x + 1875 x ) (3 + 5 x )
6.3 Integration by Parts

Let u = f (x) and v = g (x) be two differentiable functions of x. The formula for the derivative of a product of functions
is: d (u.v) = u.dv + v.du, which we rearrange as: u.dv = d (u.v) − v.du. Integrating both sides of this equation gives us:
ò u.dv = ò d ( u.v ) - ò v.du , so we have:
ò u.dv = u.v - ò v.du
We have u = f (x) Þ du = f’(x) dx, and v = g (x) Þ dv = g’(x) dx, so the final expression of the integral will be:
ò f ( x ).g ’( x )dx = f ( x ).g ( x ) - ò g ( x ).f ’( x )dx

This gives us a method of finding the integral of a product of two functions which can potentially reduce the
integral to a simpler integral.
135
www.pdfgrip.com
MATLAB incorporates the command maple(‘intparts’) which allows you to find an integral directly by parts.
This command belongs to the student library, which has to be previously loaded into the memory with the command
maple (‘with(student)’). The syntax is as follows:
maple (‘intparts (int (f (x), x), u (x))’)
Finds the integral of f(x) by parts, where u(x) is a factor of f(x) which
will be differentiated when calculating the integral.
As an example, we calculate the integral of the function (3x2 + 2x − 7)cos (x):

>> pretty(simple(sym(maple('intparts(int((3*x^2+2*x-7)*cos(x),x),3*x^2+2*x-7)'))))

2
3 x sin(x) - 13 sin(x) + 6 cos(x) x + 2 cos(x) + 2 x sin(x)

Exercise 6-5
ò ( 5x - 3 ) 4 3 x +1 dx , òe x cos ( x ) dx , ò sin [ln( x )] x 2 dx .

2
For the first integral we put u(x) = 5x 2−3:

>> pretty(simple(sym(maple('intparts(int((5*x^2-3)*4^(3*x+1),x),5*x^2-3)'))))

x 2 2 2
64 (5 - 54 log(2) - 30 x log(2) + 90 x log(2) )
1/27 --------------------------------------------------
3
log(2)

For the second integral we put u(x) = e x:

>> pretty(simple(sym(maple('intparts(int(exp(x)*cos(x),x),exp(x))'))))

1/2 exp (x) (cos (x) + sin (x))

For the third integral we put u(x) = x:

>> pretty(simple(sym(maple('intparts(int(sin(ln(x))*x^2,x),x)'))))

3 2
x (- 1 + 6 tan(1/2 log(x)) + tan(1/2 log(x)) )
1/10 -----------------------------------------------
2
1 + tan(1/2 log(x))
136
www.pdfgrip.com
6.4 Integration by Reduction and Cyclic Integration

Integration by reduction (reduction formulas) is used to integrate functions involving large integer exponents. It
reduces the integral to a similar integral where the value of the exponent has been reduced. Repeating this procedure
we obtain the value of the original integral.
The usual procedure is to perform integration by parts. This will lead to the sum of an integrated part and an
integral of a similar form to the original, but with a reduced exponent.
Cyclic integration is similar except we end up with the same integral that we had at the beginning, except for
constants. The resulting equation can be rearranged to give the original integral.
In both cases the problem lies in the proper choice of the function u(x) in the integration by parts.
MATLAB directly calculates the value of this type of integral in the majority of cases. In the worst case, the final
value of the integral can be found after one to three applications of integration by parts.
Exercise 6-6
ò sin
13
( x )cos15 ( x ) dx , òx 7 (m + nx )1/2 dx , òe 12 x cos11 ( x ) dx .
We solve the first integral by parts by putting u (x) = (sin (x))12:

>> pretty(simple(sym(maple('intparts(int(sin(x)^13*cos(x)^15,x),sin(x)^12)'))))

16 12 10 8 6
- 1/48048 cos(x) (1716 sin(x) + 792 sin(x) + 330 sin(x) + 120 sin(x)

4 2
+ 36 sin(x) + 8 sin(x) + 1)

We solve the second integral by putting u(x) = x 7:

>> pretty(simple(sym(maple('intparts(int(x^7*sqrt(m+n*x),x),x^7)'))))

3/2 7 7 7 4 3 3 5 2 2
2/109395 (m + n x) (-2048 m + 6435 n x + 4480 m n x - 3840 m n x

6 2 5 5 3 4 4 6 6 / 8
+ 3072 m n x + 5544 m n x - 5040 m n x - 6006 m n x ) / n
/

137
www.pdfgrip.com
We solve the third integral by putting u (x) = (cos (x))10:

>> pretty(simple(sym(maple('intparts(int(exp(12*x)*cos(x)^11,x),cos(x)^10)'))))

11 10
1/63918719475 exp(12 x) (2894432580 cos(x) + 2653229865 cos(x) sin(x)

9 8 7
+ 1415055928 cos(x) + 1061291946 cos(x) sin(x) + 527896512 cos(x)

6 5 4
+ 307939632 cos(x) sin(x) + 131193216 cos(x) + 54663840 cos(x) sin(x)

3 2
+ 17149440 cos(x) + 4287360 cos(x) sin(x) + 709632 cos(x) + 59136 sin(x))
6.5 Definite Integrals

The definite integral acquires its strength when it comes to applying the techniques of integration to practical
problems. Here we present some of the most common applications.
6.6 Curve Arc Length

One of the most common applications of integral calculus is to find lengths of arcs of a curve.
For a planar curve with equation y = f (x), the arc length of the curve between points with coordinates x = a and
x = b is given by the expression:
b
L=ò 1 + f ¢( x )2 dx
a
For a planar curve with parametric coordinates x = x (t) and y = y (t), the arc length of the curve between the
points corresponding to the values t = t0 and t = t1 of the parameter is given by the expression:
t1
L=ò x ¢(t )2 + y ¢(t )2 dt
t0
For a curve in polar coordinates with equation r = f (a), the arc length of the curve between the points
corresponding to the parameter values a = t0 and a = t1 is given by the expression:
a1
L=ò r 2 + r ¢(a )2 dr
a0
For a space curve with parametric coordinates x = x (t), y = y (t), z = z (t), the arc length of the curve between the
points corresponding to the parameter values t = t0 and t = t1 is given by the expression:
t1
L=ò x ¢(t )2 + y ¢(t )2 + z ¢(t )2 dt
t0
For a space curve in cylindrical coordinates with equations x = r · cos(a), y = r · sin (a), z = z, the arc length of the
curve between the points corresponding to the parameter values a = a0 and a = a1 is given by the expression:
a1
L=ò r 2 + r ¢2 + z ¢2 dr
a0
138
www.pdfgrip.com
For a space curve in spherical coordinates with equations given by x = r · sin(a) · cos(b), y = r · sin(a) · sin(b),
z = r · cos(a), the arc length of the curve between the points corresponding to the values a = a0 and a = a1 is given by
the expression:
a1
L=ò dr 2 + r 2 da 2 + r 2 sin 2 (a )db 2 dr 2
a0
Exercise 6-7
An electrical cable hangs between two towers which are 80 meters apart. The cable adopts the shape of a
catenary whose equation is:
x
y = 100 cosh
100
Calculate the arc length of the cable between the two towers.
Given the symmetry of the curve with respect to the y-axis, the limits of integration will be –40 and 40:

>> f ='100*cosh(x/100)';
>> diff(f,'x')

ans =

sinh(1/100*x)

>> pretty(simple(int('(1+sinh(1/100*x)^2)^(1/2)','x','-40','40')))

100 exp (2/5) - 100 exp(-2/5)

If we want the approximate result, we use the command numeric:

>> numeric(simple(int('(1+sinh(1/100*x)^2)^(1/2)','x','-40','40')))

ans =

82.1505

Result: 82.1505 m.
139
www.pdfgrip.com
Exercise 6-8
Calculate the arc length of the space curve represented by the parametric equations:
2
4 t
x = t y = t 3/2 z = 6
3 2
between t = 0 and t = 2.

>> pretty(simple(int('(diff(t,t)^2+diff((4/3)*t^(3/2),t)^2+diff((1/2)*t^2,t)^2)^(1/2)',
't','0','2')))
1/2 1/2
2 13 - 3/2 log(13 + 4) - 1 + 3/2 log(3)

Now we approximate the result:

>> numeric(simple(int('(diff(t,t)^2+diff((4/3)*t^(3/2),t)^2+diff((1/2)*t^2,t)^2)^(1/2)',
't','0','2')))

ans =

4.8157

Result: 4.8157.
Exercise 6-9
Find the arc length for values from a = 0 to a = 2pi of the cardioid given in polar coordinates by r = 3-3cos(a).

>> r = '3-3*cos(a)';

>> diff(r,'a')

ans =

3 * sin (a)

>> R = simple (int ('((3-3 * cos (a)) ^ 2 + (3 * sin (a)) ^ 2) ^(1/2) ',' a ', ' 0','2 * pi'))

R=

24
Result: 24.
140
www.pdfgrip.com
6.7 The Area Enclosed between Curves

The area enclosed between a curve with equation y = f (x) and the x-axis is given, in general, by the integral:
b
S= ò a
f ( x ) dx
where x = a and x = b are the abscissas of the end points of the curve.
If the curve is given in parametric coordinates x = x (t) and y = y(t), the area is given by the integral:
b
S= òa
y (t ) x ¢(t ) dt
where values of the parameter t = a and t = b correspond to the end points of the curve.
If the curve is given in polar coordinates r = f (a), the area is given by the integral:
1 a1
2 òa0
S= f (a )2 da
for the parameter values a = a0 and a = a1 corresponding to the end points of the curve.
To calculate the area between two curves with equations y = f(x) and y = g(x), we use the integral:
b
S = ò f ( x ) - g ( x ) dx
a
where x = a and x = b are the abscissas of the end points of the two curves.
When calculating these areas it is very important to take into account the sign of the functions involved since the
integral of a negative portion of a curve will be negative. One must divide the region of integration so that positive and
negative values are not computed simultaneously. For the negative parts one takes the modulus.
Exercise 6-10
Find the area of the region bounded by the curves defined below:
f ( x ) = 2 - x 2 and g ( x ) = x .
We will have to find the abscissas of the points where the curves meet. Consider the equation:
2 - x2 = x

>> solve('2-x^2=x')

ans =

[-2]
[1]

Now, we can already find the area that we have to calculate:

>> int('2-x^2-x','x',-2,1)

ans =

9/2

141
www.pdfgrip.com
Result: 4.5 square units:
The graph in Figure 6-1 illustrates the area that we had to calculate:

>> fplot('[2-x^2,x]',[-2,1])

Figure 6-1.
Exercise 6-11
Calculate the area under the normal curve between the limits -1.96 and 1.96.
2
e x /2
1.96
We need to calculate the integral ò-1.96 2p dx .

numeric(int('exp(-x^2/2)/(2*pi)^(1/2)','x',-1.96,1.96))

ans =

0.9500
142
www.pdfgrip.com
Exercise 6-12
Calculate the area of the ellipse with major and minor axes of length a and b.
2 2
y
As is well known, the equation of this ellipse is: x 2 + 2 = 1
a b b 2
Rearranging this equation for y and assuming the positive root, we obtain y = a - x2 .
a
By the symmetry of the ellipse, the area will be four times the integral of this expression between 0 and a.
a 4b a 2 - x 2
Therefore we calculate: ò
0 a
dx .

>> pretty(int('(4*b*(a^2-x^2)^(1/2))/a','x',0,'a'))

pi a b

The requested result is therefore pab.
Exercise 6-13
Calculate the length of the perimeter and the area enclosed by the curves:
cos ( t )[2 - cos(2t )] sin ( t )[2 + cos(2t )]
x (t ) = , y (t ) = .
4 4
First, in order to get an idea of what the problem is, we graphically represent the curve (see Figure 6-2):

>> x = (0:.1:2*pi);
>> t = (0:.1:2*pi);
>> x = cos(t).*(2-cos(2*t))./4;
>> y = sin(t).*(2+cos(2*t))./4;
>> plot(x,y)

143
www.pdfgrip.com
Figure 6-2.
We see that as the parameter ranges between 0 and 2p the curve described is closed. We can then calculate its
length and the area that it encloses.

>> A = simple (diff ('cos (t) * (2-cos(2*t)) / 4'))

A =

-3/8 * sin (t) + 3/8 * sin(3*t)

>> B = simple (diff (' sin (t) * (2 + cos(2*t)) / 4'))

B =

3/8 * cos (t) + 3/8 * cos(3*t)

Now we apply the formula for the length of a curve in parametric coordinates (considering only half a quadrant):

>> int ('sqrt ((-3/8 * sin (t) + 3/8 * sin(3*t)) ^ 2 + (3/8 * cos (t) + 3/8 * cos(3*t)) ^ 2)',
't', 0, pi/4)

ans =

3/8

144
www.pdfgrip.com
In total there are 8 semi-quadrants, so we have that L = 8 * (3/8) = 3.

The length of the curve is 3 units. To calculate the area, we divide it into four equal pieces (to prevent positive and
negative cancellation), where t varies between 0 and 2p. The area of each piece will be the absolute value of the
integral between 0 and 2p of the product y(t) * x’(t):

>> S = simple(int('(sin(t)*(2+cos(2*t))/4)*(-3/8*sin(t)+3/8*sin(3*t))','t',0,pi/2))

S =

-3/128 * pi

Thus, the area enclosed by the curve is 3p/32 square units.
Exercise 6-14
Calculate the length and the area enclosed by each of the following curves given in polar coordinates:
r = cos(2a ) and r = sin ( 2a ) .
We begin by representing both curves on the same graph (see Figure 6-3).

>> a = 0:.1:2*pi;
>> subplot(1,2,1)
>> r = sqrt(cos(2*a));
>> polar(a,r)
>> title(' r = (cos(2a))1/2')
>> subplot (1,2,2)
>> r = sin(2*a);
>> polar (a, r)
>> title(' r = sin(2a)')

145
www.pdfgrip.com
Figure 6-3.
Now, we calculate the arc lengths and the areas enclosed by both curves. The first curve repeats the structure for
a between 0 and p/4 four times, and the second for a between 0 and p/2:

>> r = 'sqrt(cos(2*a))';
>> A = simple(diff(r,'a'))

A =

-1/cos(2*a) ^(1/2) * sin(2*a)

We calculate the area S1 enclosed by the first curve. The area under the curve in the first quadrant is:

>> I = simple(int('(sqrt(cos(2*a)))^2','a',0,pi/4))

I=

1/2

146
www.pdfgrip.com
We have that S 1 = 4 * (1/2) * I = 1 square units.

Now, we find the arc length L1 of the first curve:

>> integral = simple (sym ('sqrt ((sqrt (cos(2*a)) ^ 2 + (-1/cos(2*a) ^(1/2) * sin(2*a)) ^ 2))'))

integral =

1/cos(2*a)^(1/2)

>> I = numeric(int('1/cos(2*a)^(1/2)','a',0,pi/4))

I =

1.3110

Then, L1 = 4 * I = 4 * 1.3110 = 5.244 units.
Now, we find the area enclosed by the second curve:

>> r = 'sin(2*a)';
>> B = simple (diff (r))

B =

2 * cos(2*a)

>> I = simple (int ('sin(2*a) ^ 2', 'a', 0, pi/2))

I =

1/4 * pi

The area will be S2 = 4 * (1/2) * I = p /2 square units.
Lastly we find the arc length L2 of the second curve:

>> I = numeric(int('sqrt(sin(2*a)^2+(2*cos(2*a))^2)','a',0,pi/2))

I =

2.4221

We have that L2 = 4 * 2.4221 = 9.6884 units.
147
www.pdfgrip.com
Exercise 6-15
Calculate the area enclosed by the curve y = sin(x) and y = cos(x) for x varying between 0 and 2p.
Figure 6-4 shows the graph of both curves on the same axis, so we can see the points of intersection and the
region between the curves.

>> fplot ('[sin (x), cos (x)], [0, 2 * pi])

Figure 6-4.
We need to know the coordinates of the points of intersection of the two curves. To do this, we solve the system
formed by their two defining equations. The command solve does not solve the problem. Therefore, since the
graph indicates that there are solutions in the intervals (0,1) and (1,2p), we will use the command maple(‘fsolve’)
to find these solutions:

>> maple ('fsolve (sin (x) = cos (x), x, 0.. 1)')

ans =

.7853981633974483

>> maple ('fsolve (sin (x) = cos (x), x, 1.. 2 * pi)')

ans =

3.926990816987242

148
www.pdfgrip.com
Now, we calculate the area between the two curves as follows:

>> S1=numeric(int('cos(x)-sin(x)','x',0,0.785398))

S1 =

0.4142

>> S2=numeric(int('cos(x)-sin(x)','x',0.785398,3.92699))

S2 =

-2.8284

>> S3=numeric(int('cos(x)-sin(x)','x',3.92699,2*pi))

S3 =

2.4142

We have I = S 1 + abs (S 2) + S 3 = 5.6569 square units.
6.8 Surfaces of Revolution

The area of the surface generated by rotating the curve with equation y = f (x) around the x-axis is given by the integral:
b
S = 2p ò f ( x ) 1 + f ’( x )2 dx
a
where x = a and x = b are the x-coordinates of the end points of the revolving curve. If the curve is given in parametric
coordinates x = x(t) and y = y(t) and the end points of the curve correspond to t = t0 and t = t1 then the rotational surface
area is given by the integral:
t1
S = 2p ò y (t ) x ’(t ) + y ’( x )2 dt
t0
The area of the surface generated by rotating the curve with equation x = f (y) around the y axis is given by
the integral:
b
S = 2p ò f ( y ) 1 + f ’( y )2 dy
a
where y = a and y = b are the y-coordinates of the end points of the rotating curve.
149
www.pdfgrip.com
Exercise 6-16
Calculate the area of the surface generated by rotating the cubic y = 12x − 9x 2 + 2x 3 around the OX axis, for x
between 0 and 5/2.
The surface of revolution has equation y 2 + z 2 = (12x − 9x 2 + 2x 3)2, and for the purposes of producing the graph
shown in Figure 6-5, we can parameterize the surface as follows:
x = t , y = cos ( u ) (12t - 9t 2 + 2t 3 ) , z = sin(u)(12t - 9t 2 + 2t 3 )

>> t = (0:.1:5/2);
>> u = (0:.5:2*pi);
>> x = ones (size (u))'* t;
>> y = cos (u)'*(12*t-9*t.^2+2*t.^3);
>> z (u) = sin(u)'*(12*t-9*t.^2+2*t.^3);
>> surf (x, y, z)

Figure 6-5.

Now we find the surface of revolution by integration:

>> I = numeric(int('(2*x^3-9*x^2+12*x)*sqrt(1+(6*x^2-18*x+12)^2)','x',0,5/2))

I =

0.8656

Thus the surface of revolution has area 2p(0.8656) = 1.7312p square units.
150
www.pdfgrip.com
Exercise 6-17
Calculate the area of the surface of revolution given by rotating the parametric curve defined by x (t ) = t − sin(t )
and y (t ) = 1 − cos(t ) around the OX axis, for t between 0 and 2p .

>> x = 't - sin (t)';
>> y = '1 - cos (t)';
>> [diff(x),diff(y)]

ans =

[1-cos (t), sin (t)]

Now, we find the requested surface area by integration:

>> I = numeric(int('(1-cos(t))*sqrt(abs((1-cos(t))^2+sin(t)^2))','t',0,2*pi))

I =

10.6667

The area will be 2p(10.6667) = 21.3334p square units.
6.9 Volumes of Revolution

The volume generated by rotating the curve with equation y = f (x) around the x-axis is given by the integral:
b
V = p ò f ( x )2 dx
a
where x = a and x = b are the x-coordinates of the end points of the rotating curve.
The volume generated by rotating the curve with equation x = f (y) around the y-axis is given by the integral:
b
V = p ò f ( y )2 dy
a
where y = a and y = b are the y-coordinates of the end points of the rotating curve.
If one cuts a volume by planes parallel to one of the three coordinate planes (for example, the plane z = 0) and if
the equation S(z) of the curve given by the cross section is given in terms of the distance of the plane from the origin
(in this case, z) then the volume is given by:
z2
V = ò S (z ) dz
z1
151
www.pdfgrip.com
Exercise 6-18
x2 y2
Calculate the volume generated by rotating the ellipse + = 1 around the OX axis and around the OY axis.
4 9
We depict half of each of the generated surfaces of revolution next to each other in Figure 6-6. The final volume
will therefore be twice that of the represented figure. The equation of the surface of revolution around the
æ x2 ö
OX axis is y 2 + z 2 = 9 ç 1 - ÷ , which can be parameterized as follows:
è 4 ø
1 1
æ t 2 ö2 æ t 2 ö2
x = t , y = 3 cos ( u ) ç 1 - ÷ , z = 3 sin(u) ç 1 - ÷
è 4ø è 4ø
Figure 6-6.
æ y2 ö
The equation of the surface of revolution around the OY axis is x 2 + z 2 = 4 ç 1 - ÷ , which can be parameterized
è 9 ø
as follows:
1 1
æ t 2 ö2 æ t 2 ö2
x = 2 cos ( u ) ç 1 - ÷ , y = t , z = 2 sin(u) ç 1 - ÷
è 9ø è 9ø

>> t = (0:.1:2);
>> u = (0:.5:2*pi);
>> x = ones (size (u))'* t;
>> y = cos(u)'*3*(1-t.^2/4).^(1/2);
>> z = sin(u)'*3*(1-t.^2/4).^(1/2);
>> subplot(1,2,1)
>> surf(x,y,z)
152
www.pdfgrip.com
>> subplot(1,2,2)
>> x = cos(u)'*2*(1-t.^2/9).^(1/2);
>> y = ones (size (u))'* t;
>> z = sin(u)'*2*(1-t.^2/9).^(1/2);
>> surf (x, y, z)

Now we calculate the generated volumes by integration:

>> V1 = int('pi*9*(1-x^2/4)','x',-2,2)

V1 =

24*pi

>> V2 = int('pi*4*(1-y^2/9)','y',-3,3)

V2 =

16*pi
Exercise 6-19
Calculate the volume generated by rotating the curve given in polar coordinates by r = 1 + cos(a) about
the OX axis.
The curve is given in polar coordinates, but there is no problem in calculating the values of x(a) and y(a) needed to
implement the volume formula in cartesian coordinates. We have:

x (a) = r (a) cos (a) = (1 + cos (a)) cos (a)

y (a) = r (a) sin (a) = (1 + cos (a)) sin (a)

>> x ='(1 + cos (a)) * cos (a)';
y ='(1 + cos (a)) * sin (a)';
>> A = simple(diff(x))

A =

-sin(2*a) - sin (a)

>> I = int ('((1 + cos (a)) * sin (a)) ^ 2 * (- sin(2*a) - sin (a))', 'a', 0, pi)

I =

-8/3

The required volume is p (abs (I)) = 8p/3 cubic units.
153
www.pdfgrip.com
6.10 Curvilinear Integrals

Let F be a continuous vector field in R3 and c: [a, b] → R3 be a continuous differentiable curve in R3. We define
ò F • ds , and call it the integral of F along the curve c, as follows:
c
ò F • ds = ò F [c(t )] • c¢(t ) dt
b
c a
Exercise 6-20
Consider the curve c (t ) = [sin (t ) ,cos (t ) ,t ] with 0 < t < 2p , and the vector field F ( x , y , z ) = xi + y j + zk .
Calculate òc F × ds.

derivC = [diff ('sin (t)'), diff('cos (t)'), diff ('t')]

derivC =

[cos (t),-sin(t), 1]

>> pretty (int ('dotprod (vector ([sin (t), cos (t), t]), vector ([cos (t),-sin(t), 1]))
', ', 0, 2 * pi))
2
2 pi
Exercise 6-21
Consider the curve c defined by the parametric equations:

7p
x = cos3 ( a ) , y = sin 3 ( a ) , z = a , 0 < a <
2
Calculate ò [sin ( z ) dx + cos ( z ) dy -
c
xydz ] . 3

derivC = [diff ('cos (a) ^ 3'), diff (' sin (a) ^ 3'), diff('a')]

derivC =

[- 3 * cos (a) ^ 2 * sin (a) 3 * sin (a) ^ 2 * cos (a) 1]

>> pretty (sym (numeric (int ('dotprod (vector ([sin (a), cos (a) - sin (a) ^ 3 *
cos (a) ^ 3]), vector ([- 3 * cos (a) ^ 2 * sin (a), 3 * sin (a) ^ 2 * cos (a), 1]))
^(1/3)', 'a', 0, 7 * pi/2))))

1/2
5/2 + 3i
154
www.pdfgrip.com
Exercise 6-22
Let the curve c be the circle of radius a: x 2 + y 2 = a 2. Calculate the value of ò
c
x 3dy - y 3dx .
First, the circle can be parameterized (x = a cos(t ), y = a sin(t )) and we can then calculate the integral as follows:

>> pretty(int('a*cos(t)^3*diff(a*sin(t),t)-a*sin(t)^3*diff(a*cos(t),t)','t',0,2*pi))

2
3/2 a pi
6.11 Approximate Numerical Integration

If it is not possible to solve an integral algebraically then MATLAB offers several techniques for numerical
approximation, including such classical algorithms as the Newton-Cotes, Simpson, Chebyshev and trapezoidal
methods. Opening the Maple student library (with the command with(student)), the available functions are the
following (all must be preceded with the command maple):
evalf (int(expr,var=#..#) uses the Chebychev method (or if that fails, the Newton-Cotes
method) to approximate the given integral
readlib(‘evalf/int’): ‘evalf/int’(expr,var=#..#) approximates the given integral using the
Newton-Cotes method
simpson(expr,var=a... b) evaluates the integral of the expression using Simpson’s
method with 4 equal size intervals
simpson(expr,var=a..b, n) evaluates the integral of the expression using Simpson’s
method with n equal size intervals
trapezoid(expr,var=a..b) evaluates the integral of the expression using the trapezoid
method with 4 equal size intervals
trapezoid(expr,var=a..b, n) evaluates the integral of the expression using the trapezoid
method with n equal size intervals
As an example, we calculate the integral of xkLn (x) between 1 and 3 using the Simpson and trapezoid methods
with 4 equal intervals.

>> maple('simpson(x^k*ln(x), x=1..3)')

ans =

1/6*3^k*log(3)+2/3*Sum((1/2+i)^k*log(1/2+i),i = 1 .. 2)+1/3*Sum((1+i)^k*log(1+i),i = 1 .. 1)

155
www.pdfgrip.com
This expression can be written as:
2 æ 2 æ 1 ö æ 1 öö 1 æ 1
k
1 k ö
3 ln(3) + ç å ç +i ÷ ln ç + i ÷ ÷ + ç å (1+i ) ln (1 + i ) ÷
k
6 3 çè i =1 è 2 ø è 2 ø ÷ø 3 è i =1 ø

>> maple('trapezoid(x^k*ln(x), x = 1..3)')

ans =

1/2*Sum((1+1/2*i)^k*log(1+1/2*i),i = 1 .. 3)+1/4*3^k*log(3)

This expression can be written as:
1 æ 3 æ 1 ö æ 1 öö 1 k
k
çç å ç 1 + i ÷ ln ç 1 + i ÷ ÷÷ + 3 ln(3)
2 è i =1 è 2 ø è 2 ø ø 4
6.12 Improper Integrals

MATLAB works with improper integrals in the same way as it works with any other type of definite integral. We will
not discuss theoretical issues concerning the convergence of improper integrals here, but within the class of improper
integrals we will distinguish two types:
1. Integrals with infinite limits: the domain of definition of the integrand is a half-line
[(a,∞) or (-∞, a)] or the entire line (-∞,∞).
2. Integrals of discontinuous functions: the given function is continuous in an interval [a, b]
except at finitely many so-called isolated singularities.
Complicated combinations of these two cases may also occur. One can also generalize this to the more general
setting of Stieltjes integrals, but to discuss this would require a course in mathematical analysis.
Exercise 6-23
Study and calculate the value of the following integrals:

-x 3/2
b dx p /2 ¥e sin( x ) ¥ x
ò , ò tan ( x ) dx , ò dx , ò dx .
x 0 1+ x2
0
x 0 0
The first of the integrals presents a singularity at x = 0, so there may be problems in a neighborhood of 0.
We have:

>> pretty(sym(limit(int(1/sqrt(x),x,a,b),a,0)))

1/2
2 b

>> pretty(simple(sym(int(1/sqrt(x),x,0,b))))

1/2
2 b

156
www.pdfgrip.com
Thus the first integral is convergent and has the value 2 b .

The second integral presents a unique singularity at x = p/2, and is divergent because:

>> pretty(sym(limit(int(tan(x),x,0,pi/2-b),b,0)))

inf

The third integral is convergent, since it can be calculated as:

>> pretty(simple(sym(int(exp(-x)*sin(x)/x,x,0,inf))))

1/4 pi

The fourth integral is divergent, because:

>> pretty(sym(limit(int(x^(3/2)/(1+x^2),x,1,b),b,inf)))

NaN
6.13 Parameter Dependent Integrals b

Consider the function of the variable y: òa f ( x , y ) dx = F ( y ) defined in the interval c £ y £ d, where the function f (x, y)
is continuous in the rectangle [a, b] × [c, d] with continuous partial derivative with respect to y in the rectangle, then
for all y such that c £ y £ d we have:
d é b b d
ò f ( x , y ) dx ù = ò f ( x , y ) dx
ê
dy ë a ú
û a dy
This result is very important, because it allows us to differentiate an integral by differentiating under the
integral sign.
Integrals dependent on a parameter can also be improper, and in addition the limits of integration may also
depend on a parameter.
If the limits of integration depend on a parameter, we have the following:
d é b( y ) b ( y ) df ( x , y )
ò f ( x , y ) dx ù = ò dx + b¢ ( y ) × f éëb ( y ) , y ùû - a¢( y ) × f [a ( y ) , y ]
ê
dy ë a ( y ) ú
û a (y) dy
provided a(y) and b(y) are defined on the interval [c, d] and have continuous derivatives a ’ (y) and b ’ (y), and the
curves a(y) and b(y) are contained in the rectangle [a, b] × [c, d].
b
Also, if the function ò f ( x , y ) dx = F ( y ) is defined on the interval [c, d] and f (x, y) is continuous in the rectangle
a
[a, b] × [c, d], then the following holds:
d b b d
ò ò f ( x , y ) dxdy =ò ò f ( x , y ) dydx
c a a c
i.e. integration under the integral sign is valid, and the order in which the integrals are evaluated can be reversed
without affecting the result.
157
www.pdfgrip.com
Exercise 6-24
Solve the following integrals using a parameter a > 0:

2
¥ tan -1 (ax ) ¥ 1-e
-x
ò0 x (1 + x 2 )
dx , ò0 x 2 dx.
For the first integral, we will start by integrating the derivative of the integrand with respect to the parameter a,
which will be easier to integrate. Once this integral is found we integrate with respect to a to find the original
integral.

>> maple('assume(a>0)');
>> pretty(simple(sym((int(diff(atan(a*x)/(x*(1+x^2)),a),x,0,inf)))))

pi
--------
2 a + 2

Now we integrate this function with respect to the variable a, to find the original integral:

>> pretty(simple(sym(int(pi/(2*a+2),a))))

1/2 log(2 a + 2) pi

To solve the second integral, we consider the following, using a as a parameter:
¥ 2
ó 1 - e - ax
f (a ) = ôô dx
õ0 x2
As in the first integral, we differentiate the integrand with respect to the parameter a, find the integral, and then
integrate with respect to a. The desired integral is then given by setting a = 1.

>> Ia = simple (sym (int (diff ((1-exp(-a*x^2)) / x ^ 2, a), x, 0, inf)))

Ia =

1/2/a ^(1/2) * pi ^(1/2)

>> s = simple(sym(int(Ia,a)))

s =

a ^(1/2) * pi ^(1/2)

Putting a = 1, we obtain the value p for the original integral.
158
www.pdfgrip.com
6.14 The Riemann Integral

The definite integral of a function f(x) gives the total area between the graph of the function and the x-axis contained
between the two limits of integration, where areas under the x-axis are regarded as negative.
Riemann defined this integral as a certain limit. Given a partition of the domain of the function into intervals
one can define two associated values, the upper sum and the lower sum. The upper sum is the sum of the areas of the
rectangles with bases the intervals of the partition and heights given by the maximum of the function in the interval.
The lower sum is similarly defined except the height of the rectangle is determined by the minimum of the function
in the interval. Thus there are infinitely many upper sums and infinitely many lower sums. One can define the upper
integral to be the infimum of the upper sums and the lower integral to be the supremum of the lower sums. If these
two values are equal, then the function is Riemann integrable and the common value is its integral.
By dividing the domain into a large number of intervals with small widths one can approximate the integral.
This method can be used when the integrand cannot be integrated algebraically.
If the function y = f (x) we have to integrate is positive and continuous on the interval of integration [a, b],
the upper and lower sums can be calculated with the following MATLAB expressions:
suminf = (b-a)/n *symsum(abs(subs(f,a + k*(b-a)/n)),k,0,n-1)
sumsup = (b-a)/n*symsum(abs(subs(f,a + k*(b-a)/n)),k,1,n)
If the function we have to integrate is negative in certain subsets of its domain, we take the modulus of the
function for the purpose of computing the integral and finding areas. One can apply this directly to the function
without having to determine the subintervals on which it is negative.
In addition, MATLAB implements the command rsums whose syntax is as follows:
rsums (f) approximates the integral of f using Riemann sums and produces a
graphical representation of rectangles
Using the maple command, loading the student library into the memory, the following commands are available:
leftsum(expr,variable=a..b)
approximates the integral of expr with respect to the variable ranging
between a and b by Riemann sums, taking 4 rectangles of equal base
whose heights are given by the value of the function expr at the left
end points of each base interval. The result is given in the form of sum
which can be evaluated using the command value (exact symbolic
evaluation), or evalf (numerical approximation).
leftsum(expr,variable=a..b, n)
same as the above except with n rectangles
rightsum(expr,var=a..b)
same as leftsum except the heights of the 4 rectangles are determined
by the right end points of each base interval
rightsum(expr,var=a..b, n)
middlesum(expr,var=a..b)
same as leftsum except the heights of the 4 rectangles are determined
by the midpoints of each base interval
middlesum(expr,var=a..b, n)
159
www.pdfgrip.com
Exercise 6-25
Approximate the value of the following integrals:

dx sin( x )
sin ( x 2 ) dx ,
1 5 7 5
ò cos ( sin( x )) dx , ò
0 2 ò
2 ln( x )
, ò
1 x
dx .
To approximate the integrals, we partition the range of integration into n intervals, find the middlesum, and then
find the limit of these sums as n tends to infinity:

>> maple('with(student)')
>> pretty(sym(maple('evalf(limit(middlesum(cos(sin(x)),x=0..1,n),n=infinity))')))

0.8687400396

> evalf(limit(middlesum(sin(x)^2,x=2..5,n),n=infinity));

1.446804

> evalf(limit(middlesum(1/log(x),x=2..7,n),n=infinity));

3.711887986

> evalf(limit(middlesum(sin(x)/x,x=1..5,n),n=infinity));

0.6038481746

Alternative methods can be used. For example, to approximate the first integral, we can calculate the limit as the
number of rectangles n tends to infinity of both the upper and lower sums. If these two limits match, then the
function is integrable and the common value is the integral.

>> maple ('f: = x - > cos (sin (x))');
>> numeric(maple('limit((1-0)/n*sum(abs(f(0+(1-0)*k/n)),k=1..n),n=infinity)'))

ans =

0.8687

>> numeric(maple('limit((1-0)/n*sum(abs(f(0+(1-0)*k/n)),k=0..n-1),n=infinity)'))

ans =

0.8687

If we calculate the integral directly, we see that the values coincide:

>> numeric (int ('cos (sin (x))', 'x', 0, 1))

ans =

0.8687

160
www.pdfgrip.com
We can produce a graphical representation of the rectangles as follows (see Figure 6-7):

>> syms x, f = cos (sin (x))
>> rsums (f)

Figure 6-7.
We perform the same steps for the second integral:

>> maple ('g: = x - > sin (x) ^ 2');
>> numeric(maple('limit((5-2)/n*sum(g(2+(5-2)*k/n),k=0..n-1),n=infinity)'))

ans =
1.4468

>> numeric(maple('limit((5-2)/n*sum(g(2+(5-2)*k/n),k=1..n),n=infinity)'))

ans =
1.4468

Since the values coincide, this common value must be the integral. We confirm this:

>> numeric(int('sin(x)^2','x',2,5))

ans =
1.4468

161
www.pdfgrip.com
To approximate the third integral we use 1,000 rectangles:

>> maple('h:=x->1/log(x)');
>> numeric('(7-2)/1000*sum(abs(h(2+(7-2)*k/1000)),k=1..1000)')

ans =
3.7096

>> numeric('(7-2)/1000*sum(abs(h(2+(7-2)*k/1000)),k=0..999)')

ans =
3.7142

The exact integral is therefore a number between 3.7096 and 3.7142, as shown by:

>> numeric(int('1/log(x)','x',2,7))

ans =
3.7119

To calculate the fourth integral we use 500 rectangles as follows:

>> maple('with(student)') ;
>> maple('evalf(middlesum(sin(x)/x,x=1..5,500))')

ans =
.60384711788763591362042282119684

This value matches that obtained by integrating directly using:

>> numeric(int('sin(x)/x','x',1,5))

ans =
0.6038
162
www.pdfgrip.com
Chapter 7
Integration in Several Variables

and Applications
In this chapter we will describe how to solve multivariate integrals with MATLAB, and give applications of double
and triple integrals to calculate areas and volumes. We also give some other typical applications of multivariate
integral calculus.
The relevant MATLAB functions have already been introduced in the previous chapter. They are the following:
syms x, int(f(x), x) or int(‘f(x)’, ‘x’)
int (int (‘f(x,y)’, ‘x’), ‘y’)
syms x and int (int (f(x,y), x), y)
int (int (int (… int (‘f(x,y…z)’, ‘x’), ‘y’)…), ‘z’)
Computes ò ò ¼ ò f ( x , y ,¼, z ) dxdy ¼dz .
syms x y z, int(int(int(…int(f(x,y…z), x), y)…), z)
syms x a b, int (f (x), x, a, b)
b
Computes the definite integral ò a
f ( x )dx
int (‘f (x)’, ‘x’, ‘a’, ‘b’)
b
Computes the definite integral ò a
f ( x )dx
int(int(‘f(x,y)’, ‘x’, ‘a’, ‘b’), ‘y’, ‘c’, ‘d’))
b d
a c
f ( x , y )dxdy
syms x y b c d, int (int (f(x,y), x, a, b), y, c, d)
b d
a c
f ( x , y )dxdy
163
www.pdfgrip.com
Chapter 7 ■ Integration in Several Variables and Applications
int(int(int(…int(‘f(x,y,…,z)’, ‘x’, ‘a’, ‘b’), ‘y’, ‘c’, ‘d’),…), ‘z’, ‘e’, ‘f’)
b d f
Computes ò ò ¼ò f ( x , y ,¼ , z ) dxdy ¼ dz
a c e
syms x y z a b c d e f, int (int (int (… int (f(x,y,…,z), x, a, b), y, c, d),…), z, e, f)

b d f
a c e
In addition, with prior use of the command maple, the following commands can be used for integration:
int (f (x), x)
int (f (x), x = a…b)
b
f ( x )dx
int(f(x), x=a..b, continuous)
Computes the definite integral removing the check for continuity of
the function in the interval (a, b)
int (int (f(x,y), x), y)
Computes the indefinite integral òò f ( x , y )dxdy
int(int(int(…int(f(x,y,…,z), x), y),…), z)
int (int (f(x,y), x = a..b, y = c…d))
b d
a c
f ( x , y )dxdy
int (int (int (… int (f(x,y,…,z), x = a…b, y = c…d,…, z = e…f))))
b d f
a c e
Commands for multivariate integration are also available in the student package, again requiring the prior use of
the command maple. Their syntax is as follows:
Doubleint (expr, var1, var2, name)
Finds the inert double integral of the expression in the variables var1
and var2, where the domain of integration is specified by name
Doubleint (expr, var1=a1..b1, var2=a2..b2)
Finds the specified definite inert double integral
Tripleint (expr, var1, var2, var3, name)
Finds the specified indefinite inert triple integral
Tripleint (expr, var1=a1..b1, var2 = a2…b2, var3 = a3…b3)
Finds the specified inert definite triple integral
164
www.pdfgrip.com
changevar (expr1=expr2,Fnc,var)
Changes the variable in the expression Fnc (Int, Limit, or Sum). The
new variable is defined in terms of the old where expr1 is an expression
in the old variable (appearing in Fnc) and expr2 is an expression in the
new variable var. Used to change variables in integrals.
changevar ({eqn1,eqn2},Doubleint,[var1,var2])
Change of variable in double integrals. The equations defining the new
variables in terms of the old are in the form expr1i = expr2i (i = 1, 2),
where the expr1i are expressions in the old variables, and expr2i are
expressions in the new variables var1 and var2.
changevar ({eqn1,eqn2,eqn3},Tripleint,[var1,var2,var3])
Change of variables for triple integrals
completesquare (expression)
Completes the square with respect to all variables appearing in the
specified polynomial expression
completesquare (expr, variable)
Completes the square of the specified polynomial expression with
respect to the given variable
integrand (expression)
Extracts all integrands of an expression containing integrals (Int,
Doubleint, Tripleint)
powsubs (expr1=expr2,expr)
Replaces all occurrences of expr1 in expr by expr2 (expr1 is a
subexpression of expr)
powsubs ({eqn1,…,eqnn},expr) or subs([eqn1,…,eqnn],expr)
Simultaneously performs the substitutions in expr described by the n
equations eqni of the form expr1i = expr2i, i = 1… n
7.1 Areas and Double Integrals

If we consider an enclosure S, we can find its area through the use of double integrals. If the area S is determined by
boundary curves whose equations are given in cartesian coordinates, its area A is given by the formula:
A = òò dx dy
S
If, for example, S is determined by a < x < b and f (x) < y < g(x), the area will be:
b g (x)
A = ò dx ò dy
a f (x)
165
www.pdfgrip.com
If S is determined by h (a) < x < k (b) and c < y < d, the area will be:
d h(b)
A = ò dy ò dx
c h(a )
If the region S is determined by curves whose equations are given in polar coordinates with radius vector r and
angle a, its area A is given by the formula:
A = òò r da dr
S
If, for example, S is determined by s < a < t and f(a) < r < g(a) then
t g (a )
A = ò da ò r dr
s f (a )
Exercise 7-1
Calculate the area of the region above the OX axis bounded by the OX axis, the parabola y 2 = 4x and the line
x + y = 3.
First, we create a graphical representation of the problem, which is presented in Figure 7-1:

>> fplot ('[(4*x) ^(1/2), 3-x]', [0,4,0,4])

Figure 7-1.
166
www.pdfgrip.com
We see that in the enclosed region y is limited between 0 and 2 (0 < y < 2) and x is limited between the curves
x = ( y ^ 2)/4 and x = 3 − y. We calculate the requested area as follows:

>> int(int('1','x','y^2/4','3-y'),'y',0,2)

ans =

10/3
Exercise 7-2
Calculate the area in the first quadrant bounded between the semicubical parabola y 2 = x 3 and the bisector of the
first quadrant.
First, we graphically represent the problem (see Figure 7-2):

>> fplot ('[^(3/2) x, x]', [-4.5-1-4])

Figure 7-2.
167
www.pdfgrip.com
Inspecting the graph we see that the enclosed area is defined for x ranging between 0 and 1 (0 < x < 1) and y
ranging between the curves y = x and y = x ^( 3/2). Therefore the area of the requested region can be calculated as:

>> int(int('1','y','x^(3/2)','x'),'x',0,1)

ans =

1/10
Exercise 7-3
Calculate the area outside the circle with polar equation r = 2, and inside the cardioid with polar equation
r = 2 (1 + cos (a )).

>> a = 0:0. 1:2 * pi;
>> r = 2 * (1 + cos (a));
>> polar (a, r)
>> hold on;
>> r=2*ones(size(a));
>> polar (a, r)

Figure 7-3.
168
www.pdfgrip.com
Inspecting the graph, we see that, by symmetry, we can calculate half of the required area by varying a between
0 and Pi / 2 (0 < a < Pi /2) and r between the curves r = 2 and r = 2 (1 + cos (a)):

>> pretty(int(int('r','r',2,'2*(1+cos(a))'),'a',0,pi/2))

1/2 pi + 4

The required area is therefore p + 8 square units.
7.2 Surface Area by Double Integration

The surface area of a surface S defined by z = f(x, y), where (x, y) ranges over a region R in the OXY plane, is given by:
2 2
æ¶z ö æ¶z ö
S = òò 1+ ç ÷ +ç ÷ dx dy
R
è¶x ø è¶ y ø
The surface area of a surface S defined by x = f(y, z), where (y, z) ranges over a region R in the OYZ plane,
is given by:
2 2
æ¶x ö æ¶x ö
S = òò 1+ ç ÷ +ç ÷ dy dz
R
è¶y ø è¶z ø
The surface area of a surface S defined by y = f (x, z), where (x, z) ranges over a region R in the OXZ plane, is
given by:
2 2
æ¶ y ö æ¶ y ö
S = òò 1+ ç ÷ +ç ÷ dx dz
R
è¶x ø è¶z ø
Exercise 7-4
Calculate the area of the surface of the cone x 2 + y 2 = z 2, limited above the OXY plane and cut by the cylinder
x 2 + y 2 = b y.
The projection of the surface onto the OXY plane is the disc bounded by the circle with equation x 2+ y 2 = by.
Then we can find the surface area via the first formula above as follows:

>> maple('z:=(x,y)->sqrt((x^2+y^2)/a)');
>> pretty(int(int('(1+diff(z(x,y),x)^2+diff(z(x,y),y)^2)^(1/2)','x','-(b*y-y^2)^(1/2)',
'(b*y-y^2)^(1/2)'),'y',0,'b'))
2 1/2
b pi (a + 1)
1/4 -------------
1/2
a
169
www.pdfgrip.com
Exercise 7-5
Calculate the area of the paraboloid x 2 + y 2 = 2 z limited below the plane z = 2.

>> maple('z:=(x,y)->(x^2+y^2)/2');
>> pretty(maple('(1+diff(z(x,y),x)^2+diff(z(x,y),y)^2)^(1/2)'))

2 2 1/2
(1 + x + y )

This expression is suitable for a transformation to polar coordinates:

>> maple('m:=(x,y)->sqrt(1+x^2+y^2)');
>> pretty(simple('m(r*cos(a),r*sin(a))'))

2 1/2
(1 + r )

The requested integral is calculated as 4 times the integral delimited by the first quadrant of the circle r = 2 :

>> pretty(simple(int(int('r*sqrt(1+r^2)','r',0,2),'a',0,pi/2)))

1/2
1/6 (5 5 - 1) pi

( )
The result of the integral will be 4 times the previous value, i.e. 2 5 5 - 1 p/ 3 .
Figure 7-4 shows the graphical representation of the problem:

>> [x, y] = meshgrid(-3:.1:3);
>> z=(1/2)*(x.^2+y.^2);
>> mesh(x,y,z)
>> hold on;
>> z=2*ones(size(z));
>> mesh(x,y,z)
>> view(-10,10)

170
www.pdfgrip.com
Figure 7-4.
7.3 Volume Calculation by Double Integrals

The volume V of a cylindroid limited in its upper part by the surface with equation z = f(x, y), its lower part by the OXY
plane and laterally by the straight cylindrical surface which cuts the OXY plane bordering a region R, is:
V = òò f ( x , y ) dx dy = òò z dx dy
R R
The volume V of a cylindroid limited in its upper part by the surface with equation x = f(y, z), its lower part by the
OYZ plane and laterally by the straight cylindrical surface which cuts the OYZ plane bordering a region R, is:
V = òò f ( y , z ) dy dz = òò x dy dz
R R
The volume V of a cylindroid limited in its upper part by the surface with equation y = f(x, z), its lower part by the
OXZ plane and laterally by the straight cylindrical surface which cuts the OXZ plane bordering a region R, is:
V = òò f ( x , z ) dx dz = òò y dx dz
R R
171
www.pdfgrip.com
Exercise 7-6
Calculate the volume in the first octant bounded between the OXY plane, the plane z = x + y + 2 and the cylinder
x 2 + y 2 = 16.
We begin by graphically representing the problem (see Figure 7-5) with the plane in cartesian coordinates and
parameterizing the cylinder:

>> t =(0:.1:2*pi);
>> u =(0:.1:10);
>> x = 4 * cos (t)'* ones (size (u));
>> y = 4 * sin (t)'* ones (size (u));
>> z = ones (size (t))'* u;
>> mesh (x, y, z)
>> hold on;
>> [x,y]=meshgrid(-4:.1:4);
>> z = x + y + 2;
>> mesh (x, y, z)
>> set(gca,'Box','on');
>> view(15,45)

Figure 7-5.
172
www.pdfgrip.com
The required volume is then given by the integral:

>> pretty(simple(int(int('x+y+2','y',0,'sqrt(16-x^2)'),'x',0,4)))

128/3 + 8 pi
Exercise 7-7
Calculate the volume bounded by the paraboloid x 2 + 4 y 2 = z and laterally by the cylinders with equations y 2 = x
and x 2 = y.

>> [x,y]=meshgrid(-1/2:.02:1/2,-1/4:.01:1/4);
>> z = x ^ 2 + 4 * y. ^ 2;
>> mesh(x,y,z)
>> hold on;
>> y=x.^2;
>> mesh(x,y,z)
>> hold on;
>> x=y.^2;
>> mesh(x,y,z)
>> set(gca,'Box','on')
>> view(-60,40)

Figure 7-6.
173
www.pdfgrip.com
The volume is calculated via the following integral:

>> pretty(int(int('x^2+4*y^2','y','x^2','sqrt(x)'),'x',0,1))

3/7
7.4 Volume Calculation and Triple Integrals

The volume of a three-dimensional region R whose boundary surface equations are expressed in cartesian
coordinates is given by the triple integral:
òòò dx dy dz
R
The volume of a three-dimensional region R whose boundary surface equations are expressed in cylindrical
òòò r dz dr da
R
The volume of a three-dimensional region R whose boundary surfaces equations are expressed in spherical
òòò r
2
sin ( b ) drdbda
R
Exercise 7-8
Calculate the volume bounded by the paraboloid x 2 + y 2 = z and the cylinder with equation z = a 2 − y 2.
The volume will be four times the following integral:

>> pretty(simple(int(int(int('1','z','a*x^2+y^2','a^2-y^2'),'y',0,
'sqrt((a^2-a*x^2)/2)'),'x',0,'sqrt(a)')))

4
a (log(-a) - log(a))
1/8 ---------------------
1/2
(-2 a)

174
www.pdfgrip.com
We can assume a > 0 and simplify the value of 4 times the previous integral as follows (note this gives the real
part of the given expression):

>> pretty(simple(4*int(int(int('1','z','a*x^2+y^2','a^2-y^2'),'y',0,
'sqrt((a^2-a*x^2)/2)'),'x',0,'sqrt(a)')))

7/2 1/2
1/4 a 2 pi
Exercise 7-9
Calculate the volume bounded by the cylinders z = x 2 and 4 − y 2 = z.


>> [x, y] = meshgrid(-2:.1:2);
>> z = x ^ 2;
>> mesh(x,y,z)
>> hold on;
>> z = 4 - y. ^ 2;
>> mesh (x, y, z)

Figure 7-7.
175
www.pdfgrip.com
The volume will be four times the following integral:

>> pretty(simple(int(int(int('1','z','x^2','4-y^2'),'y',0,'sqrt(4-x^2)'),'x',0,2)))

2 pi

The final result for the volme is therefore 8p.
Exercise 7-10
Calculate the volume enclosed by the cylinder r = 4cos(a), the plane z = 0 and the sphere with equation r 2 + z 2 = 16.
The volume of the enclosure, given in cylindrical coordinates, is:

>> pretty(simple(int(int(int('r','z',0,'sqrt(16-r^2)'),'r',0,'4*cos(a)'),'a',0,pi)))

256/9 + 64/3 pi
Exercise 7-11
Calculate the volume enclosed by the cone b = p /4 and the sphere with equation r = 2 k cos (b).
The volume of the enclosure, given in spherical coordinates, will be four times the result of the following integral:

>> pretty(simple(int( int(int('r^2*sin(b)','r',0,'2*k*cos(b)'),'b',0,pi/4),'a',0,pi/2)))

3
1/4 k pi

The final result is therefore k3p.
7.5 Green’s Theorem

Let C be a simple closed planar curve, and R the region consisting of C and its interior. If f and g are continuous
functions with continuous first partial derivatives in an open set D containing R, then:
æ¶n ¶m ö
ò m( x , y ) dx + n( x , y ) dy = òò çè ¶ x - ¶ y ÷ø dA
C R
176
www.pdfgrip.com
Exercise 7-12
Calculate the following integral using Green’s theorem:
ò (x +e ) dx + (2 y + cos(x ))dy
y
C
where C is the boundary of the region enclosed by the parabolas y = x 2 , x = y 2.
The two parabolas intersect at the points (0,0) and (1,1). We graphically represent the problem (see Figure 7-8)
and calculate the integral:

plot([x^2,sqrt(x)],x,0,1.2);
Figure 7-8.
>> maple ('m:=(x,y) - > x+exp (sqrt (y))');

>> maple ('n:=(x,y) - > 2 * y + cos (x)');
>> maple ('evalf (int (int (diff (n(x,y), x)-diff (m(x,y), y), y = x ^ 2.. sqrt (x)),
x = 0.. 1))')

ans =
-.67644120041679251512532326651204

177
www.pdfgrip.com
The integral of the form can also be calculated as follows:

>> numeric(int(int('diff(n(x,y),x)-diff(m(x,y),y)','y','x^2','sqrt(x)'),'x',0,1))

ans =

-0.6764
7.6 The Divergence Theorem

Let Q be a domain with the property that each straight line passing through a point inside a domain cuts its border
at exactly two points. In addition, suppose the boundary S of the domain Q is a closed oriented surface with exterior
normal vector n. If f is a vector field that has continuous partial derivatives on Q, then:
òòS
f • n dS = òòò Div( f ) dV
Q
The left-hand side of this equality is called the outflow of the vector field f across the surface S.
Exercise 7-13
Use the divergence theorem to find the outflow of the vector field f = (xy + x 2, yz + xy 2, z z, xz + xyz 2 ) across the
surface of the cube in the first octant bounded by the planes x = 2, y = 2 and z = 2.

>> maple ('f: = vector([x*y+x^2*z*y, y*z+x*y^2*z, x*z+x*y*z^2])');
>> maple ('v: = vector ([x, y, z])');
>> int(int(int('diverge(f,v)','x',0,2),'y',0,2),'z',0,2)

ans =

72
7.7 Stokes’ Theorem

Let S be an oriented surface of finite area defined by a function f(x,y), having boundary C with unit normal n. Suppose
F is a continuous vector field defined on S such that the component functions of F have continuous partial derivatives
at each non-boundary point of S. Then:
ò F × dr = ò ò ( curl F ) × n ds
C S
178
www.pdfgrip.com
Exercise 7-14
Use Stokes’ theorem to evaluate the line integral:
òC
- y 3dx + x 3dy - z 3dz
where C is the intersection of the cylinder x 2 + y 2 = 1 and the plane x + y + z = 1, and the orientation of C
corresponds to counterclockwise rotation of the OXY plane.
The curve C bounds the surface S defined by z = 1 − x − y = f (x,y) for (x, y) in the domain D = {(x,y) / x 2 + y 2 = 1}.
We put F = − y 3 i + x 3 j − z 3 k.
Now, we calculate the curl of F and integrate over the surface S:

>> maple ('F: = vector([-y^3,x^3,z^3])');
>> maple ('v: = vector ([x, y, z])');
>> pretty(sym(maple('curl(F,v)')))

2 2
[0, 0, 3 x + 3 y]

Therefore, we have to calculate the integral ∫D (3x 2 + 3y 2 ) dx dy. Changing to polar coordinates, we obtain:

>> pretty(simple(int(int('3*r^3', 'a',0,2*pi),'r',0,1)))

3/2 pi
179
www.pdfgrip.com
Chapter 8
Differential Equations
Although it implements only a relatively small number of commands related to this topic, MATLAB’s treatment of
differential equations is nevertheless very efficient. We shall see how we can use these commands to solve each type
of differential equation algebraically. Numerical methods for the approximate solution of equations and systems of
equations are also implemented.
The basic command used to solve differential equations is dsolve. This command finds symbolic solutions
of ordinary differential equations and systems of ordinary differential equations. The equations are specified by
symbolic expressions where the letter D is used to denote differentiation, or D2, D3, etc, to denote differentiation of
order 2,3,..., etc. The letter preceded by D (or D2, etc) is the dependent variable (which is usually y), and any letter
that is not preceded by D (or D2, etc) is a candidate for the independent variable. If the independent variable is not
specified, it is taken to be x by default. If x is specified as the dependent variable, then the independent variable is t.
That is, x is the independent variable by default, unless it is declared as the dependent variable, in which case the
independent variable is understood to be t.
You can specify initial conditions using additional equations, which take the form y(a) = b or Dy(a) = b,..., etc.
If the initial conditions are not specified, the solutions of the differential equations will contain constants of integration,
C1, C2,..., etc. The most important MATLAB commands that solve differential equations are the following:
dsolve( ‘equation’, ‘v’)
Solves the given differential equation, where v is the independent variable (if ‘v’ is
not specified, the independent variable is x by default). This returns only explicit
solutions.
dsolve( ‘equation’, ‘condition_initial’,..., ‘v’)
Solves the given differential equation subject to the specified initial condition
dsolve( ‘equation’, ‘cond1’, ‘cond2’,..., ‘condn’, ‘v’)
Solves the given differential equation subject to the specified initial conditions
dsolve( ‘equation’, ‘cond1,cond2,...,condn’, ‘v’)
Solves the given differential equation subject to the specified initial conditions
dsolve( ‘eq1’, ‘eq2’,..., ‘eqn’, ‘cond1’, ‘cond2’,..., ‘condn’, ‘v’)
Solves the given differential system subject to the specified initial conditions
(explicit)
dsolve( ‘eq1,eq2,...,eqn’, ‘cond1,cond2,...,condn’ , ‘v’)
Solves the given differential system subject to the specified initial conditions
181
www.pdfgrip.com
Chapter 8 ■ Differential Equations
There is another group of important MATLAB commands that solve differential equations, requiring the prior use
of the command maple. They are as follows:
dsolve (eqn, fnc (var))
Symbolically solves the ordinary differential equation eqn for the function
fnc(var). The solution is usually returned in implicit form as an equation in var,
fnc (var) and the implied constants C_1,..., C_n. var is an independent variable
and fnc is the dependent variable. The result may also contain calls to the DEsol
command.
dsolve (expr, fnc (var))
Solves the differential equation expr = 0
dsolve({deqn,cond1,...,condn},fnc(var))
Solves the differential equation deqn subject to the initial conditions
cond1,..., condn
dsolve (deqn, fnc (var), explicit = true)
Gives the solution in explicit form of the differential equation deqn, if possible
dsolve(deqn,fnc(var),method=laplace)
Forces the use of the Laplace transform method in the solution of the differential
equation
dsolve(deqn,fnc(var),type=series)
Forces the use of the Taylor series method in the solution of the differential
equation, with the degree of the polynomial solution determined by the global
variable Order
dsolve(deqn,fnc(var),output=basis)
The solution is given in terms of the specified basis
powseries[powsolve](deqn,cond1,...,condn)
Solves the differential equation deqn subject to the initial conditions given in
terms of power series
dsolve ({deqn1,..., deqnn}, {(var) fnc1... fncn (var)})
Solves the specified system of differential equations for the independent variable
var and the dependent variables fnc1,..., fncn
dsolve ({deqn1,..., deqnn, cond1,..., condm}, {(var) fnc1..)fncn (var)})
Solves the specified system of differential equations for the independent variable
var and the dependent variables fnc1,..., fncn, subject to the initial conditions
cond1,..., condm
DEsol (deqn, fnc (var))
Gives the solution of the differential equation deqn for the function fnc(var) in
inert form
182
www.pdfgrip.com
Examples are given below.

First, we solve differential equations of first order and first degree, both with and without initial values.

>> pretty(dsolve('Dy = a*y'))

exp(a t) C1

>> pretty(dsolve('Df = f + sin(t)'))

-1/2 cos (t) - 1/2 sin (t) + exp (t) C1

The previous two equations can also be solved in the following way:

>> pretty (sym (maple ('dsolve (diff (y (x), x) = a * y, y (x))')))

y (x) = exp(a x) _C1

>> pretty(sym(maple('dsolve(diff(f(t),t)=f+sin(t),f(t))')))

f(t) = - 1/2 cos(t) - 1/2 sin(t) + exp(t) _C1

>> pretty(dsolve('Dy = a*y', 'y(0) = b'))

exp(a t) b

>> pretty(dsolve('Df = f + sin(t)', 'f(pi/2) = 0'))

exp (t)
-1/2 cos (t) - 1/2 sin (t) + 1/2 - -----------
exp(1/2 pi)

Now, we solve an equation of second degree and first order:

>> y = dsolve ('(Dy) ^ 2 + y ^ 2 = 1', ' (0) = 0', 'y(s)')

y =

[sin (s)]
[-sin (s)]

We can also solve this in the following way:

>> pretty(maple('dsolve({diff(y(s),s)^2 + y(s)^2 = 1, y(0) = 0}, y(s))'))

y (s) = sin (s), y (s) = - sin (s)

Now we solve an equation of second order and first degree:

>> pretty (dsolve ('D2y = - a ^ 2 * y ', 'y (0) = 1, Dy (pi/a) = 0'))

cos (a t)

183
www.pdfgrip.com
Next we solve a couple of systems, both with and without initial values:

>> pretty(dsolve('Dx = y', 'Dy = -x'))

x (t) = (t) C1 + C2 cos (t), y (t) = C1 cos (t) - C2 sin (t)

>> S=dsolve('Df = 3*f+4*g', 'Dg = -4*f+3*g');[S.f, S.g]

ans =

[exp(3*t)*cos(4*t)*C1+exp(3*t)*sin(4*t)*C2, -exp(3*t)*sin(4*t)*C1+ exp(3*t)*cos(4*t)*C2]

>> T= dsolve('Df = 3*f+4*g, Dg = -4*f+3*g', 'f(0)=0, g(0)=1') ;[T.f, T.g]

ans =

[ exp(3*t)*sin(4*t), exp(3*t)*cos(4*t)]

This last system can also be solved in the following way:

>> pretty(maple('dsolve({diff(f(x),x)= 3*f(x)+4*g(x), diff(g(x),x)=- 4*f(x) + 3*g(x),
f(0)=0,g(0)=1}, {f(x),g(x)})'))

{f (x) = exp (3 x) sin(4 x), g (x) = exp(3 x) cos(4 x)}

We now solve a differential equation by the Laplace transform method:

>> dsolve ({diff (y (t), t$ 2) + 5 * diff (y (t), t) + 6 * y (t) = 0, y (y) (0) (0) = 0, D = 1},
y (t), method = laplace);

y (t) = - exp(-3 t) + exp(-2 t)

Next we solve a differential system in series form:

>> dsolve({diff(y(x),x)=z(x)-y(x)-x,diff(z(x),x)=y(x),y(0)=0,z(0)=1},{y(x), z(x)}, type=series);

2 3 4 5 6
{y(x) = x - x + 1/2 x - 5/24 x + 1/15 x + O(x ),

2 3 4 5 6
z(x) = 1 + 1/2 x - 1/3 x + 1/8 x - 1/24 x + O(x )}
8.1 Separation of Variables

A differential equation is said to have separable variables if it can be written in the form f (x) dx = g(y) dy.
This type of equation can be solved immediately by putting ∫ f (x) dx = ∫ g(y) dy + C.
If MATLAB cannot directly solve a differential equation with the function dsolve, then we can try to express it
in the above form and solve the given integrals algebraically, which does not present particular difficulties for the
program, given its versatility in symbolic computation.
184
www.pdfgrip.com
Exercise 8-1
Solve the differential equation:

y cos(x ) dx −(1 + y 2) dy = 0, y (0) = 1.
First of all we try to solve it directly. The equation can be written in the form:
cos ( x ) y ( x )
y¢ ( x ) = .
1 + y ( x )2

>> dsolve('Dy = y * cos(x) /(1+y^2)')

ans =
0
exp(C33 + t*cos(x))*exp(-wrightOmega(2*C33 + 2*t*cos(x))/2)

Thus the differential equation appears not to be solvable with dsolve. However, in this case, the variables are
separable, so we can solve the equation as follows:

>> pretty(solve('int(cos(x), x) = int((1+y^2) / y, y)'))

+- -+
| / 2 \ |
| | y | |
| asin| log(y) + -- | |
| \ 2 / |
| |
| / 2 \ |
| | y | |
| pi - asin| log(y) + -- | |
| \ 2 / |
+- -+

Thus, after a little rearrangement, we see that the general solution is given by:
sin(x) = log(y ) + 1/2 y 2 + C.
We now find the value of the constant C via the initial condition, putting x = 0 and y = 1.

>> C = simple('solve(subs(x = 0, y = 1, sin(x) = log(y) + 1/2 * y ^ 2 + C), C)')

C =

-1/2

Thus the final solution is sin(x) = log(y) + 1/2 y 2 − ½.
In the same way you can solve any other differential equation with separable variables.
The above differential equation is also solvable directly by using:

>> pretty(maple('dsolve(diff(y(x), x) = y(x) * cos(x) /(1 + y(x) ^ 2), y(x))'))

2
log(y(x)) + 1/2 y(x) - sin(x) = _C1

185
www.pdfgrip.com
8.2 Homogeneous Differential Equations

Consider a general differential equation of first degree and first order of the form
M(x,y) dx = N(x,y) dy.
This equation is said to be homogeneous of degree n if the functions M and N satisfy:
M(tx, ty) = t n M(x,y),
N(tx, ty) = t n N(x,y).
For this type of equation, we can transform the initial differential equation (with variables x and y), via the change
of variable x = vy, into another (separable) equation (with variables v and y). The new equation is solved by separation
of variables and then the solution of the original equation is found by reversing the change of variable.
Exercise 8-2

(x 2 − y 2) dx + x y dy = 0.
First we check if the equation is homogeneous:

>> maple('m:=(x,y)->x^2-y^2');
>> maple('n:=(x,y)->x*y');
>> pretty(sym(maple('collect(m(t*x,t*y),t)')))

2 2 2
(x - y ) t

>> pretty(sym(maple('collect(n(t*x,t*y),t)')))

2
t y x

Thus the equation is homogeneous of degree 2. To solve it we apply the change of variable x = vy.
Before performing the change of variable, it is advisable to load the library difforms, using the command
maple(‘with(difforms)’), which will allow you to work with differential forms. Once this library is loaded it is
also convenient to use the command maple(‘defform(v=0,x=0,y=0)’), which allows you to declare all variables
which will not be constants or parameters in the differentiation.

>> maple('with(difforms)');
>> maple('defform(v=0,x=0,y=0)');

Now we can make the change of variable x = vy, and group the terms in d(v) and d(y).

>> pretty(simple(sym(maple('subs(x=v*y,m(x,y)*d(x)+n(x,y)*d(y))'))))

2 2 2 2
(v y - y ) (y d(v) + v d(y)) + y v d(y)

186
www.pdfgrip.com
>> pretty(sym(maple('collect(v^2*y^3*d(v)+v^3*y^2*d(y)-y^3*d(v),{d(v),d(y)})')))

3 2 2 3 3
v y d(y) + (v y - y ) d(v)

If we divide the previous expression by v 3y 3, and group the terms in d (v) and d (y), we already have an equation in
separated variables.

>> pretty(sym(maple('collect(((v^2*y^3-y^3)*d(v)+v^3*y^2*d(y))/(v^3*y^3), {d(v),d(y)})')))

2 3 3
(v y - y ) d(v) d(y)
----------------- + ----
3 3 y
v y

The previous expression can be simplified.

>> pretty(maple('convert(collect(((v^2*y^3-y^3)*d(v)+v^3*y^2*d(y))/(v^3*y^3),
{d(v),d(y)}),parfrac,y)'))

2
(v - 1) d(v) d(y)
------------- + ----
3 y
v

Now, we solve the equation:

>> pretty(simple(sym('int((v^2-1)/v^3,v)+int(1/y,y)')))

1
log(v) + ---- + log(y)
2
2 v

Finally we reverse the change of variable:

>> pretty(simple(sym('subs(v=x/y,log(v)+1/2/v^2+log(y))')))

2
y
log(x) + 1/2 ----
2
x

Thus the general solution of the original differential equation is:
1 y2
log ( x ) + = C.
2 x2
187
www.pdfgrip.com
Now we can represent the solutions of this differential equation graphically. To do this we graph the solutions with
parameter C, which is equivalent to the following contour plot of the function defined by the left-hand side of the
above general solution (see Figure 8-1):

>> [x,y]=meshgrid(0.1:0.05:1/2,-1:0.05:1);
>> z=y.^2./(2*x.^2)+log(x);
>> contour(z,65)
Figure 8-1.
8.3 Exact Differential Equations

The differential equation
M(x,y) dx + N(x,y) dy = 0
is said to be exact if ∂N/∂x = ∂M/∂y. If the equation is exact, then there exists a function F such that its total differential
dF coincides with the left-hand side of the above equation, i.e.:
dF = M(x,y) dx + N(x,y) dy
therefore the family of solutions is given by F(x,y) = C.
The exercise below follows the usual steps of an algebraic solution to this type of equation.
188
www.pdfgrip.com
Exercise 8-3

(−1 + y e xy + y cos(xy )) dx + (1 + x e xy + x cos(xy )) dy = 0.
First, we try to solve the equation with dsolve:

>> maple ('m:=(x,y) - > - 1 + y * exp(x*y) + y * cos(x*y)');
>> maple ('n:=(x,y) - > 1 + x * exp(x*y) + x * cos(x*y)');
>> dsolve ('m(x,y) + n(x,y) * Dy = 0')

ans =

[empty sym]

Thus the function dsolve does not give a solution to the proposed equation. We are going to try to solve the
equation using the classical algebraic method.
First we check that the proposed differential equation is exact.

>> pretty(simple(diff('m(x,y)','y')))

(y exp(y x) - y sin(y x)) x + exp(y x) + cos(y x)

>> pretty(simple(diff('n(x,y)','x')))

(y exp(y x) - y sin(y x)) x + exp(y x) + cos(y x)

Since the equation is exact, we can find the solution in the following way:

>> solution1 = maple ('simplify (int (m(x,y), x) + g (y))')

solution1 =

-x + exp(y*x) + sin(y*x) + g (y)

Now we find the function g(y) via the following condition:
diff(int(m(x,y), x) + g(y), y) = n(x,y).

>> pretty(sym(maple('simplify(int(m(x,y),x)+g(y))')))

-x + exp(y x) + sin(y x) + g (y)

>> pretty (sym (maple ('simplify (diff (-x+exp(y*x) + sin(y*x) + g (y), y)'))))

d
exp(y x) x + x cos(y x) + --- g (y)
dy

189
www.pdfgrip.com
>> pretty (simple (sym ('solve (x * exp(y*x) + x * cos(y*x) + diff (g (y), y) = n(x,y),))))
((((diff (g (y), y))')))

ans = 1

Thus g'(y) = 1, so the final solution will be, omitting the addition of a constant:

>> pretty (simple (sym (maple ('subs (g (y) = int(1,y),-x+exp(y*x) + sin(y*x) + g (y))'))))

-x + exp(y x) + sin(y x) + y

To graphically represent the family of solutions, we draw the following contour plot of the above expression
(Figure 8-2):

>> [x,y]=meshgrid(-2*pi/3:.2:2*pi/3);
>> z =-x+exp (y.*x) + free (y.*x) + y;
>> contour(z,100)
Figure 8-2.

In the following section we will see how any reducible differential equation can be transformed to an exact
equation using an integrating factor.
190
www.pdfgrip.com
8.4 Linear Differential Equations

A linear first order differential equation is an equation of the form:
dy/dx + P(x) y = Q(x)
where P(x) and Q(x) are given functions of x.
Differential equations of this type can be transformed into exact equations by multiplying both sides of the
equation by the integrating factor:
eò
P ( x ) dx
and the general solution is then given by the expression:
æ e - ò P ( x )dx ö æ e ò P ( x )dxQ( x )dx ö .

ç ÷ç ò ÷
è øè ø
MATLAB implements these solutions of linear differential equations, and offers them whenever the integral
appearing in the integrating factor can be found.
Exercise 8-4

x dy/dx + 3 y = x sin(x ).

>> pretty (simple (dsolve ('x * Dy + 3 * y = x * sin (x)')))

1/3 x sin(x) + exp(-3 t/x) C1

Another more efficient solution of the differential equation that includes implicit solutions is as follows:

>> pretty(sym(maple('simplify(dsolve(x*diff(y(x),x)+3*y(x)=x*sin(x),y(x)))')))

3 2
x cos(x) - 3 x sin(x) + 6 sin(x) - 6 x cos(x) - _C1
y(x) = - -----------------------------------------------------
3
x
8.5 Ordinary High-Order Equations

An ordinary linear differential equation of order n has the following general form:
n
åa ( x ) y ( ) ( x ) = a ( x ) y ( x ) + a ( x ) y’( x ) + a ( x ) y" ( x ) +¼+ a ( x ) y ( ) ( x )

k =0
k
k
0 1 2 n
n
= f ( x ).
If the function f(x) is identically zero, the equation is called homogeneous. Otherwise, the equation is called
non-homogeneous. If the functions ai(x) (i = 1, ..., n) are constant, the equation is said to have constant coefficients.
191
www.pdfgrip.com
A concept of great importance in this context is that of a set of linearly independent functions. A set of functions
{f1(x), f2(x), ..., fn(x)} is linearly independent if, for any x in their common domain of definition, the Wronskian
determinant of the functions is non-zero. The Wronskian determinant of the given set of functions, at a point x of
their common domain of definition, is defined as follows:
f1 ( x ) f2 (x ) f3 (x ) ¼ fn ( x )
f1¢ ( x ) f 2¢ ( x ) f 3¢ ( x ) ¼ f n¢ ( x )
f1¢¢ ( x ) f 2¢¢ ( x ) f3 (x ) ¼
¢¢
f n¢¢ ( x ) = W (x)
¼ ¼ ¼ ¼ ¼
f1(n-1) ( x ) f 2(n-1) ( x ) f 3(n-1) ( x ) ¼ f n(n-1) ( x )
The MATLAB command maple('Wronskian') allows you to calculate the Wronskian matrix of a set of functions.
Its syntax is:
maple ('Wronskian(V,variable)')
computes the Wronskian matrix corresponding to the vector of functions V with
independent variable x.
A set S = {f1(x), ..., fn(x)} of linearly independent non-trivial solutions of a homogeneous linear equation of order n
a0 ( x ) y ( x ) + a1 ( x ) y ¢ ( x ) + a2 ( x ) y ¢¢ ( x ) +¼+ an ( x ) y (n ) ( x ) = 0
is called a set of fundamental solutions of the equation.

If the functions ai(x) (i = 1, ..., n) are continuous in an open interval I, then the homogeneous equation has a set of
fundamental solutions S = {fi(x)} in I.
In addition, the general solution of the homogeneous equation will then be given by the function:
n
f ( x ) = åci fi ( x )
i =0
where {ci} is a set of arbitrary constants.

The equation:
n
a0 + a1m + a2m 2 +¼+ anm n = åaim i = 0
i =0
is called the characteristic equation of the homogeneous differential equation with constant coefficients.
The solutions of this characteristic equation determine the general solutions of the corresponding differential
equation.
192
www.pdfgrip.com
Exercise 8-5
Show that the set of functions

{e x , xe x , x 2e x }
is linearly independent.

>> W=sym(maple('Wronskian(vector([exp(x),x*exp(x), x^2*exp(x)]),x)'))

W =
[exp(x), x*exp(x), x^2*exp(x)]
[exp(x), exp(x)+x*exp(x), 2*x*exp(x)+x^2*exp(x)]
[exp(x), 2*exp(x)+x*exp(x), 2*exp(x)+4*x*exp(x)+x^2*exp(x)]

>> pretty(determ(W))

3
2 exp(x)

This gives us the value of the Wronskian, which is obviously always non-zero. Therefore the set of functions is
linearly independent.
8.6 Higher-Order Linear Homogeneous Equations with Constant

Coefficients
The homogeneous linear differential equation of order n
n
åa ( x ) y ( ) ( x ) = a ( x ) y ( x ) + a ( x ) y¢ ( x ) + a ( x ) y ² ( x ) + ¼ + a ( x ) y ( ) ( x )
k =0
k
k
0 1 2 n
n
=0
is said to have constant coefficients if the functions ai(x) (i = 1, ..., n) are all constant (i.e. they do not depend on the
variable x).
The equation:
n
a0 + a1m + a2m 2 + ¼ + anm n = åaim i = 0
i =0
is called the characteristic equation of the above differential equation. The solutions (m1, m2, ..., mn) of this
characteristic equation determine the general solution of the associated differential equation.
If the mi (i = 1, ..., n) are all different, the general solution of the homogeneous equation with constant coefficients is:
y ( x ) = c1e m1 x + c 2e m2 x + ¼ + cn e mn x
where c1, c2, ..., cn are arbitrary constants.

If some mi is a root of multiplicity k of the characteristic equation, then it determines the following k terms of
the solution:
ci e mi x + ci +1 xe mi x + ci +2 x 2e mi x + ¼ + ci +k x k e mi x .
193
www.pdfgrip.com
If the characteristic equation has a complex root mj = a + bi , then its complex conjugate mj +1 = a − bi is also a
root. These two roots determine a pair of terms in the general solution of the homogeneous equation:
c j e ax cos ( bx ) + c j +1e ax sin ( bx ) .
MATLAB directly applies this method to obtain the solutions of homogeneous linear equations with constant
coefficients, using the command dsolve or maple('dsolve').
Exercise 8-6
Solve the following equations:

3 y ² + 2 y ¢ + 5y = 0
2 y ² + 5 y ¢ + 5 y = 0, y (0) = 0, y ¢(0) = 1 2

>> pretty(dsolve('3*D2y+2*Dy-5*y=0'))

C1 exp(t) + C2 exp(- 5/3 t)

>> pretty(dsolve('2*D2y+5*Dy+5*y=0','y(0)=0,Dy(0)=1/2'))

1/2 1/2 1/2 1/2
2/15 3 5 exp(- 5/4 t) sin(1/4 3 5 t)
Exercise 8-7
Solve the differential equation

9 y ²² - 6 y ¢¢¢ + 46 y ² - 6 y ¢ + 37 y = 0.

>> pretty (simple (dsolve('9*D4y-6*D3y+46*D2y-6*Dy+37*y=0')))

/ t \ / t \
C46 cos(t) + C47 sin(t) + C44 cos(2 t) exp| - | + C45 sin(2 t) exp| - |
\ 3 / \ 3 /

Looking at the solution, it is evident that the characteristic equation has two pairs of complex conjugate solutions.

>> solve('9*x^4-6*x^3+46*x^2-6*x+37=0')

ans =

[ i]
[ -i]
[1/3 + 2 * i]
[1/3-2 * i]

194
www.pdfgrip.com
You can also solve the equation as follows:

>> pretty (sym (maple ('dsolve (9 * diff (y (x), x$ 4) - 6 * diff (y (x), x$ 3) + 46 * diff
(y (x), x$ 2)-6 * diff (y (x), x) + 37 * y (x) = 0 y (x))')))

y (x) =

_C1 exp(1/3 x) sin(2 x) + _C2 exp(1/3 x) cos(2 x) + _C3 sin (x) + _C4 cos (x)
8.7 Non-Homogeneous Equations with Constant Coefficients.

Variation of Parameters
Consider the non-homogeneous linear equation with constant coefficients:
n
åa ( x ) y ( ) ( x ) = a ( x ) y ( x ) + a ( x ) y¢ ( x ) + a ( x ) y ² ( x ) + ¼ + a ( x ) y ( ) ( x )
k =0
k
k
0 1 2 n
n
= f ( x ).
Suppose {y1(x), y2(x),....., yn(x)} is a linearly independent set of solutions of the corresponding homogeneous
equation:
a0 ( x ) y ( x ) + a1 ( x ) y ¢ ( x ) + a2 ( x ) y ² ( x ) + ¼ + an ( x ) y (n ) ( x ) = 0.
A particular solution of the non-homogeneous equation is given by:

n
y p ( x ) = åui ( x ) yi ( x )
i =1
where the functions ui(x) are obtained as follows:
f ( x )Wi [ y1 ( x ) , y 2 ( x ) ,¼, yn ( x )]
ui ( x ) = ò dx ( i = 1,¼,n ) .
W [ y1 ( x ) , y 2 ( x ) ,¼, yn ( x )]
Here Wi[y1(x), y2(x), ..., yn(x)] is the determinant of the matrix obtained by replacing the i-th column of the
Wronskian matrix W[y1(x), y2(x), ..., yn(x)] by the transpose of the vector (0,0,..., 0, 1).
The solution of the non-homogeneous equation is then given by combining the general solution of the
homogeneous equation with the particular solution of the non-homogeneous equation. If the roots mi of the
characteristic equation of the homogeneous equation are all different, the general solution of the non-homogeneous
equation is:
y ( x ) = c1e m1 x + c 2e m2 x + ¼ + cn e mn x + y p ( x ) .
If some of the roots are repeated, we refer to the general form of the solution of a homogeneous equation
discussed earlier.
195
www.pdfgrip.com
Exercise 8-8
Solve the following differential equations:

y ² + 4 y ¢ + 13 y = x cos2 ( 3 x ) ,
y ² - 2 y ¢ + y = e x ln ( x ) .
We will follow the algebraic method of variation of parameters to solve the first equation. We first consider the
characteristic equation of the homogeneous equation to obtain a set of linearly independent solutions.

>> solve('m^2+4*m+13=0')

ans =

[- 2 + 3 * i]
[- 2 - 3 * i]

>> maple ('f: = x - > x * cos(3*x) ^ 2');
>> maple ('y1: = x - > exp(-2*x) * cos(3*x)');
>> maple ('y2: = x - > exp(-2*x) * sin(3*x)');
>> maple ('W: = x - > Wronskian ([y1 (x), y2 (x)], x)');
>> pretty(simple(sym(maple('det(W(x))'))))

3 exp(-4 x)

We see that the Wronskian is non-zero, indicating that the functions are linearly independent. Now we calculate
the functions Wi (x ), i = 1, 2.

>> maple ('W1: x-= > array ([[0, y2 (x)], [1, diff ((y2) (x), x)]])');
>> pretty(simple(sym(maple('det(W1(x))'))))

-exp(-2 x) sin (3 x)

>> maple ('W2: x-= > array ([[y1 (x), 0], [diff ((y1) (x), x), 1]])');
>> pretty(simple(sym(maple('det(W2(x))'))))

exp(-2 x) cos(3 x)

Now we calculate the particular solution of the non-homogeneous equation.

>> maple('u1:=x->factor(simplify(int(f(x)*det(W1(x))/det(W(x)),x)))');
>> maple ('u1 (x)')

ans =

1/14652300*exp(2*x)*(129285*cos(9*x)*x-6084*cos(9*x)-28730*sin(9*x)*x-
13013*sin(9*x)+281775*cos(3*x)*x-86700*cos(3*x)-187850*sin(3*x)*x-36125*sin(3*x))

196
www.pdfgrip.com
>> maple('u2:=x->factor(simplify(int(f(x)*det(W2(x))/det(W(x)),x)))');
>> maple ('u2 (x)')

ans =

1/14652300 * exp(2*x) * (563550 * cos(3*x) * x+108375 * cos(3*x) + 845325 * sin(3*x) *
x-260100 * sin(3*x) + 28730 * cos(9*x) * x+13013 * cos(9*x) + 129285 * sin(9*x) * x-6084 *
sin(9*x))

>> maple ('yp: = x - > factor (simplify (y1 (x) * (x) u1 + y2 (x) * u2 (x)))');
>> maple('yp(x)')

ans =

-23/1105 * x * cos(3*x) ^ 2 + 13436/1221025 * cos(3*x) ^ 2 + 24/1105 * cos(3*x) * sin(3*x) *
x + 3852/1221025 * cos(3*x) * sin(3*x) + 54/1105 * x-21168/1221025

Then we can write the general solution of the non-homogeneous equation:

>> maple (' y: = x - > simplify (c1 * y1 (x) + c2 * y2 (x) + yp (x))');
>> maple ('combine (y (x), trig)')

ans =

C1 * exp(-2*x) * cos(3*x) + c2 * exp(-2*x) * sin(3*x)-23/2210 * x * cos(6*x) + 1/26 * x +
6718/1221025 * cos(6*x)-2/169 + 12/1105 * x * sin(6*x) + 1926/1221025 * sin(6*x)

Now we graphically represent a set of solutions, for certain values of c1 and c2 (see Figure 8-3)

>> ezplot(simple(sym(maple('subs(c1=-5,c2=4,y(x))'))),[-1,1])
>> hold on
>> ezplot(simple(sym(maple('subs(c1=-5,c2=-4,y(x))'))),[-1,1])
>> ezplot(simple(sym(maple('subs(c1=5,c2=-1,y(x))'))),[-1,1])
>> ezplot(simple(sym(maple('subs(c1=5,c2=1,y(x))'))),[-1,1])

197
www.pdfgrip.com
Figure 8-3.
There is a MATLAB command that allows you to solve differential equations by the method of variation of
parameters directly. Its syntax is as follows:
maple ('DEtools [varparam]([expr1,...,exprn], exprrhs, varind)')
Finds the general solution by the method of variation of parameters of an ordinary
differential equation in the independent variable varind, where exprrhs is the right-hand
side of the original equation (non-homogeneous part) and the expri (i = 1... n) are the
solutions of the corresponding homogeneous equation of order n. The command value can
help to obtain the final solution.
For the second differential equation we directly apply dsolve, obtaining the solution. First, we find the general
solution of the homogeneous equation:

>> maple ('solhomog: = dsolve (diff (y (x), x$ 2) - 2 * diff (y (x), x) + y (x) = 0 y (x))')

ans =

solhomog: = y(x) = _C1 * exp (x) + _C2 * exp (x) * x

As the differential equation is of order 2 we must take two solutions (for two pairs of any values of C1 and C2).
The simplest solutions that can be taken are those which form a basis for the set of solutions (exp (x) and x exp (x)),
which can be found directly as follows:

>> maple('solhomog:=dsolve(diff(y(x),x$2)-2*diff(y(x),x)+y(x)=0,y(x),output=basis)')

ans =

solhomog := [exp(x), exp(x)*x]
198
www.pdfgrip.com
As the inhomogeneous part of the right-hand side of the equation is e x ln (x ), and the independent variable is x,
we can already apply the command varparam as follows:

>> pretty(sym(maple('DEtools[varparam](solhomog,exp(x)*ln(x),x)')))
2 2
_C[1] exp(x) + _C[2] exp(x) x + (- 1/2 x log(x) + 1/4 x ) exp(x)

+ (x log(x) - x) exp(x) x

The second differential equation is solved directly as follows:

>> pretty(simple(dsolve('D2y-2*Dy+y=exp(t)*log(t)')))

2 2
1/4 exp(t) (2 t log(t) - 3 t + 4 C1 + 4 C2 t)
8.8 Non-Homogeneous Linear Equations with Variable

Coefficients. Cauchy-Euler Equations
A non-homogeneous linear equation with variable coefficients of the form
n
åa x
k =0
k
k
y ( k ) ( x ) = a0 y ( x ) + a1 xy ¢ ( x ) + a2 x 2 y ² ( x ) +¼+ an x n y (n ) ( x )
= f (x)
is called a Cauchy–Euler equation.
This equation can be reduced to a homogeneous linear equation with constant coefficients by replacing x = et.
This leads us to solve the equation:
a0 + a1k + a2 k (k - 1) + ... + an k (k - 1)...(k - n + 1) = 0
The roots ki of multiplicity ai correspond to the solutions:

x ki , x ki ln x , x ki ln 2 x ,..., x ki lnai -1 x
The complex roots p ± qi of multiplicity a correspond to the solutions:

x p cos ( q ln x ) , x p ln x cos ( q ln x ) ,¼, x p ( ln x ) cos ( q ln x ) .
a -1
x p sin ( q ln x ) , x p ln x sin ( q ln x ) ,¼, x p ( ln x ) sin ( q ln x ) .

a -1
MATLAB solves this type of equation directly with the command dsolve or maple('dsolve').
199
www.pdfgrip.com
Exercise 8-9
Solve the following differential equation:

x 3 y ²¢ + 16 x 2 y ² + 79 xy ¢ + 125 y = 0.

>> pretty(simple(dsolve('x^3*D3y+16*x^2*D2y+79*x*Dy+125*y=0')))

C1 + C2 sin(3 log(x)) x + C3 cos(3 log(x)) x
--------------------------------------------
5
x
8.9 The Laplace Transform

Suppose f(t) is a function defined in the interval (0, ∞). The Laplace transform of f (t) is the function F(s) defined by:
¥
F ( s ) = L { f (t )}( s ) = òe - st f ( t ) dt .
0
We say that f(t) is the inverse Laplace transform of F(s), so that

L-1 {F (s )}( t ) = f (t ) .
MATLAB provides the commands maple('laplace') and maple('invlaplace') to calculate the Laplace transform
and inverse Laplace transform of an expression with respect to a variable. The syntax is as follows:
maple ('laplace (expression, t, s)')
Calculates the Laplace transform of a given expression with respect to t.
The transformed variable is s.
maple ('invlaplace (expression, s, t)')
Computes the inverse Laplace transform of the given expression with respect to s.
The inverse variable is t.

>> pretty(sym(maple('laplace(t^(3/2)-exp(t)+sinh(a*t), t, s)')))

1/2
pi 1 a
3/4 ----- - ----- + -------
5/2 s - 1 2 2
s s - a

>> pretty(sym(maple('invlaplace(s^2/(s^2+a^2)^(3/2), s, t)')))

- t BesselJ(1, a t) a + BesselJ(0, a t)

The Laplace transform and its inverse are used to solve certain differential equations. The method is to calculate
the Laplace transform of each term of the equation to obtain a new differential equation, which we then solve. Finally,
we find the solution of the original equation by applying the inverse Laplace transform to the solutions just found.
200
www.pdfgrip.com
MATLAB provides the ‘laplace’ option in the maple(‘dsolve’) command, which forces the program to solve the
differential equation using the Laplace transform method. The syntax is as follows:

maple ('dsolve (equation, func (var), 'laplace'))

Exercise 8-10
Solve the differential equation

y ² + 2 y ¢ + 4 y = x - e - x , y ( 0 ) = 1, y ¢ ( 0 ) = 1
using the Laplace transform method.

First, we calculate the Laplace transform of each side of the differential equation, and we apply the initial
conditions.

>> maple('L:=s->laplace(diff(y(x),x$2)+2*diff(y(x),x)+4*y(x),x,s)');
>> pretty(simple(sym(maple('subs(y(0)=1,(D(y))(0)=1,L(s))'))))
2
laplace(y(x), x, s) s - s - 3 + 2 laplace(y(x), x, s) s + 4 laplace(y(x), x, s)

>> maple('L1:=s->laplace(x-exp(-x),x,s)');
>> pretty(simple(sym('L1(s)')))

1 1
---- - -----
2 s + 1
S

We then solve the Laplace transformed differential equation:

>> pretty(simple(sym(maple('solve(L(s)=L1(s),laplace(y(x),x,s))'))))

4 3 2
s y(0) + (3 y(0) + D(y)(0)) s + (2 y(0) + D(y)(0) - 1) s + s + 1
-------------------------------------------------------------------
2 3 2
s (s + 3 s + 6 s + 4)

Now we substitute the given initial conditions into the solution.

>> maple('TL:=s->solve(L(s)=L1(s),laplace(y(x),x,s))');
>> pretty(simple(sym('subs(y(0)=1,(D(y))(0)=1,TL(s))')))

4 3 2
s + 4 s + 2 s + s + 1
--------------------------
2 2
s (s + 1) (s + 2 s + 4)

201
www.pdfgrip.com
This gives the solution of the Laplace transformed equation. To calculate the solution of the original equation we
calculate the inverse Laplace transform of the solution obtained in the previous step.

>> maple('TL0:=s->simplify(subs(y(0)=1,(D(y))(0)=1,TL(s)))');
>> solution=simple(sym(maple('invlaplace(TL0(s),s,x)')));
>> pretty(solution)

1/2 1/2 1/2
1 sin(3 x) 3 35 cos(3 x)
1/4 x - 1/8 - -------- + 5/8 ---------------- + ---- -----------
3 exp(x) exp(x) 24 exp(x)

This gives the solution of the original differential equation.
We could also have solved it directly via:

>> pretty(simple(sym(maple('dsolve({diff(y(x),x$2)+2*diff(y(x),x)+4*y(x)= x-exp(-x),
y(0)=1,D(y)(0)=1},y(x),laplace)'))))

1/2 1/2 1/2
1 sin(3 x) 3 35 cos(3 x)
y(x) = 1/4 x - 1/8 - -------- + 5/8 ---------------- + ---- ------------
3 exp(x) exp(x) 24 exp(x)
8.10 Systems of Linear Homogeneous Equations with Constant

Coefficients
A system of differential equations, written as X '(t) = AX(t), has a general solution of the form:
n
X ( t ) = åciVi e lit
i =1
where the eigenvalues {lk} (k = 1, 2, ... n) corresponding to the eigenvectors {Vk} of the matrix A of the system are all
assumed to be different.
If an eigenvalue lk is a complex number ak + bki, then it generates the following component of the overall solution:
C k1Wk1 e lk t + C k2Wk2 e lk t
where:
Wk1 =
1
2
( ) i
(
Vk + V k cos ( bk t ) + Vk - V k sin ( bkt ) ,
2
)
Wk2 =
i
2
( ) 1
(
Vk - V k cos ( bkt ) - Vk + V k sin ( bkt ) .
2
)
Here Vk is the eigenvector corresponding to the eigenvalue lk and V k is its conjugate.
If there is an eigenvalue li of multiplicity m > 1, then it will generate a portion of the general solution of the form:
ci e li tVi + ci +1te li tVi +1 + ci +2t 2e li tVi +2 + ××× + ci +kt k e li tVk .
MATLAB can solve this type of system directly, simply by using the command dsolve or maple('dsolve') with the
familiar syntax.
202
www.pdfgrip.com
Exercise 8-11
Solve the following system of equations:

x ¢ = -5 x + 3 y ,
y ¢ = -2 x + 10 y .

>> S = dsolve('Dx=-5*x+3*y,Dy=-2*x-10*y','t');
>> pretty(sym([S.x,S.y]))

[-2 C1 exp(-8 t) + 3 C1 exp(-7 t) + 3 C2 exp(-7 t) - 3 C2 exp(-8 t) ,

-2 C1 exp(-7 t) + 2 C1 exp(-8 t) + 3 C2 exp(-8 t) - 2 C2 exp(-7 t)]

You can also use the following syntax:

>> pretty (sym (maple ('dsolve ({diff (x (t), t) =-5 * x (t) + 3 * y (t), diff (y (t), t) =
-2 * x (t) - 10 * y (t)}, {x (t), y (t)})')))

{x(t) = -2 _C1 exp(-8 t) + 3 _C1 exp(-7 t) + 3 _C2 exp(-7 t) - 3 _C2 exp(-8 t)

y(t) = -2 _C1 exp(-7 t) + 2 _C1 exp(-8 t) + 3 _C2 exp(-8 t) - 2 _C2 exp(-7 t)}
8.11 Systems of Linear Non-Homogeneous Equations with

Constant Coefficients
Now let us consider systems of non-homogeneous differential equations with constant coefficients of the form
X ′ = A X + F(t).
The general solution of the homogeneous system X ′ = AX takes the form X = F(t )C . A particular solution of the
non-homogeneous system is:
X p = F(t )òF -1 ( t ) F ( t ) dt .
The general solution of the non-homogeneous system will be X = F(t )C + Xp, which is, using the previous
expression:
X = F ( t )C + F(t )òF -1 ( t ) F ( t ) dt .
This method is a generalization to systems of equations of the method of variation of parameters for simple
equations.
MATLAB can solve such systems of equations directly with the command dsolve or maple('dsolve'), provided the
integrals that appear in the solution can be evaluated.
203
www.pdfgrip.com
Exercise 8-12
Solve the following system of equations:

x ¢ - y ¢ = e -t
y ¢ + 5 x + 2 y = sin(3 + t )
with initial conditions x(0) = x0 and y(0) = y0.

>> S=dsolve('Dx-Dy=exp(-t),Dy+5*x+2*y=sin(3+t)','x(0)=xo,y(0)=yo','t');
>> pretty(simple(sym([S.x,S.y])))

[5/7 xo exp(-7 t) + 2/7 xo - 2/7 yo + 2/7 yo exp(-7 t) - 1/6 exp(-t)
- 5/42 exp(-7 t) + 2/7 - 7/50 exp(-7 t) sin(3) + 1/50 exp(-7 t) cos(3)
+ 7/50 sin(3 + t) - 1/50 cos(3 + t) ,

- 5/7 xo + 5/7 xo exp(-7 t) + 2/7 yo exp(-7 t) + 5/7 yo + 5/6 exp(-t) - 5/42 exp(-7 t)
- 5/7- 7/50 exp(-7 t) sin(3) + 1/50 exp(-7 t) cos(3) + 7/50 sin(3 + t)- 1/50 cos(3 + t)]
8.12 Higher Order Equations and Approximation Methods

When the known algebraic methods for solving differential equations and systems of differential equations offer no
solution, we usually resort to methods of approximation. The approximation methods can involve both symbolic
and numerical work. The symbolic approach yields approximate algebraic solutions, and its most representative
technique is the Taylor series method. The numerical approach yields a solution in the form of a finite set of solution
points, to which a curve can be fitted by various algebraic methods (interpolation, regression,...). This curve will be an
approximate solution of the differential equation. Among the most common numerical methods is the Runge–Kutta
method.
Approximation methods are most commonly employed to find the solution of equations and systems of
differential equations of order and degree greater than one, where the exact solution cannot be obtained by other
methods.
8.13 The Taylor Series Method

This method provides approximate polynomial solutions of general differential equations, and is based on the Taylor
series expansion of functions. MATLAB offers the option ‘series’ for the command maple(‘dsolve’), which allows you to
solve equations by this method. Its syntax is as follows:

maple ('dsolve (equation, func (var), 'series'))

There is also the command maple('powsolve'), which gives a power series solution of linear differential
equations, and whose syntax is as follows:

maple ('powseries [powsolve](equation, cond1, . . . , condn) ')

Using the command maple('convert(polynom)') you can convert a complicated solution to a polynomial in
powers of the variable.
204
www.pdfgrip.com
Exercise 8-13
Solve the following two equations by the Taylor series method:
4 x 2 y ¢¢ + 4 xy ¢ + ( x 2 - 1) y = 0 ,
yy '' + (y' )2 + 1 = 0 with the initial conditions y (0) = 1 and y' (0) = 1.

>> pretty(simple(sym(maple('dsolve(4*x^2*diff(y(x),x$2)+4*x*diff(y(x),x)+
+(x^2-1)*y(x)=0,y(x),series)'))))

2 4 6 6
y(x) = (_C1 x (1 - 1/24 x + 1/1920 x + O(x )) + _C2 log(x) (O(x ))

2 4 6 / 1/2
+ _C2 (1 - 1/8 x + 1/384 x + O(x ))) / x
/
>> pretty(simple(sym(maple('convert(",polynom)'))))

2 4
1/2 2 4
_C2 (1 - 1/8 x + 1/384 x )
y(x) = _C1 x (1 - 1/24 x + 1/1920 x ) + ---------------------------
1/2
x

>> pretty(simple(sym(maple('dsolve({y(x)*diff(y(x),x$2)+diff(y(x),x)^2+1=0,
y(0)=1,D(y)(0)=1},y(x),series)'))))

2 3 4 5 6
y(x) = 1 + x - x + x - 3/2 x + 5/2 x + O(x )
Exercise 8-14
Solve the following two systems of equations using the Taylor series method:
x ² + y ¢ - 4 x + 12 = 0
y ² - 10 x ¢ - y + 7 = 0
x (0) = y (0) = x ¢(0) = y ¢(0) = 1
and
x ² + 2 x ¢ + 2 y ¢ + 3z ¢ + x = 1
y¢ + z¢ - x = 0
x ¢ + z = 0.

205
www.pdfgrip.com
>> pretty(simple(sym(maple('dsolve({diff(x(t),t$2)+diff(y(t),t)-4*x+12=0,
diff(y(t),t$2)-10*diff(x(t),t)-y(t)+7=0,x(0)=1,y(0)=1,D(x)(0)=1,D(y)(0)=1},
{x(t),y(t)},series)'))))

2 3 4 147 5 6
{y(t) = 1 + t + 2 t - 89/6 t + 1/6 t + --- t + O(t ),
40

2 53 4 5 6
x(t) = 1 + t - 9/2 t + -- t - 1/30 t + O(t )}
24

>> pretty(simple(sym(maple('dsolve({diff(x(t),t$2)+2*diff(x(t),t)+2*diff(y(t),t)+ +3*diff
(z(t),t)+x(t)=1,diff(y(t),t)+diff(z(t),t)-x(t)=0,diff(x(t),t)+z(t)=0}, {x(t),y(t),z(t)},
series)'))))

2
{x(t) = x(0) + D(x)(0) t + (- D(x)(0) - 1/2 x(0) + 1/2) t

3 4
+ (1/2 D(x)(0) + 1/3 x(0) - 1/3) t + (- 1/6 D(x)(0) - 1/8 x(0) + 1/8) t

5 6
+ (1/24 D(x)(0) + 1/30 x(0) - 1/30) t + O(t ),

2 3
z(t) = z(0) + x(0) t + 1/2 D(x)(0) t + (- 1/3 D(x)(0) - 1/6 x(0) + 1/6) t

4
+ (1/8 D(x)(0) + 1/12 x(0) - 1/12) t

5 6
+ (- 1/30 D(x)(0) - 1/40 x(0) + 1/40) t + O(t ),

2 3
y(t) = y(0) + x(0) t + 1/2 D(x)(0) t + (- 1/3 D(x)(0) - 1/6 x(0) + 1/6) t

4
+ (1/8 D(x)(0) + 1/12 x(0) - 1/12) t

5 6
+ (- 1/30 D(x)(0) - 1/40 x(0) + 1/40) t + O(t ) }
8.14 The Runge–Kutta Method

The Runge–Kutta method gives a set of data points to which you can fit a curve, approximating the solution of
a differential equation. Maple provides the option numeric for the command maple('solve') which enables the
calculation of approximate numerical solutions of differential equations. Its syntax is:

maple ('dsolve (equation, func (var), 'numeric'))

206
www.pdfgrip.com
Exercise 8-15
Solve the following equation using the Runge–Kutta method:

3(y '')2 = y '''y ' with the initial conditions y '(0) = y ''(0) = 1.

>> maple ('f: = dsolve ({3 * diff (y (x), x$ 2) ^ 2 = diff (y (x), x$ 3) * diff (y (x), x),
y (0) = 1/2, D (y) (0) = 1, (D@@2) (y) (0) = 1}, y (x), numeric)');

Now, in order to graph the solution, we calculate various points of the solution function f generated above
(see Figure 8-4).
Figure 8-4.
>> [maple('f(-0.3)'),maple('f(-0.2)'),maple('f(-0.1)'),maple('f(0)'), maple('f(0.1)'),

maple('f(0.2)'),maple('f(0.3)')]

ans =

{x = -.3,y(x) = .2350889359260396}{y(x) = .3167840433732281, x = -.2} {y(x) =
.4045548849869109, x = -.1}{y(x) = .5000000000000000, x = 0} {x = .1, y(x) =
.6055728090006967}{y(x) = .7254033307597474, x = .2} {y(x) = .8675444679682489, x =
.3000000000000000}

>> y = [.2350889359260396,.3167840433732281,.4045548849869109,.5,.6055728090006967,.725403330
7597474,.8675444679682489];

>> plot ((-0.3:.1:0.3), y)
207
www.pdfgrip.com
We find the degree 2 polynomial which is the best fit to the set of solution points. The equation of this parabola
will be an approximate solution of the differential equation.

>> pretty(vpa(poly2sym(polyfit((-0.3:.1:0.3),y,2))))

2
.5747427827483573 x + 1.041293962469090 x +.4991457846921903

This yields a degree 2 polynomial approximation to the solution y(x) of the equation.
8.15 Partial Differential Equations

MATLAB implements several commands which enable you to solve partial differential equations, all of which require
the prior use of the command maple. We have:
pdesolve (pdeqn, fnc(var1,...,varn))
Solves the partial differential equation pdeqn for the function fnc(var1,..,varn)
pdesolve (exprpd, fnc(var1,...,varn))
Solves the partial differential equation exprpd = 0
DEtools[PDEchangecoords](pdeqn,[varind1,...,varindn],option)
Converts the partial differential equation pdeqn in the independent variables
varind1, varindn to the new coordinate system defined by the option. The
possible coordinate systems are bipolar, cardioid, cassinian, elliptic, hyperbolic,
invcassinian, invelliptic, logarithmic, logcosh, maxwell, parabolic, polar, rose,
and tangent. For three dimensions possible systems are bipolarcylindrical,
bispherical, cardioidal, cardioidcylindrical,casscylindrical, confocalellip,
confocalparab, conical, cylindrical, ellcylindrical, ellipsoidal, hypercylindrical,
invcasscylindrical, invellcylindrical, invoblatespheroidal, invprospheroidal,
logcoshcylindrical, logcylindrical,maxwellcylindrical, oblatespheroidal,
paraboloidal, paracylindrical, prolatespheroidal, rosecylindrical, sixsphere,
spherical, tangentcylindrical, tangentsphere, and toroidal.
DEtools [PDEchangecoords](pdeqn,[va1,...,van],option,[vn1,...,vnn])
Performs the change of coordinates specifying the new independent variables
DEtools[PDEchangecoords]({pdeqn1,...,pdeqnm},[varind1,...,varindn],option)
Performs the change of coordinates in the specified system of partial
differential equations

>> pretty(sym(maple('pdesolve( diff(f(x,y),x,x)+5*diff(f(x,y),x,y)=3, f(x,y) )')))

2
f(x, y) = 3/2 x + _F1(y) + _F2(y - 5 x)

208
www.pdfgrip.com
>> pretty(sym(maple('pdesolve( 3*diff(g(x,y),x)+7*diff(g(x,y),x,y)=x*y, g(x,y) ) ')))

2 2
g(x, y) = 1/6 x y - 7/18 x + _F1(y) + exp(- 3/7 y) _F2(x)

>> pretty(sym(maple('pdesolve( diff(h(x,y),x,x)-diff(h(x,y),y,y)=0, h(x,y)) ')))

h(x, y) = _F1(y + x) + _F2(y - x)

>> pretty(sym(maple('pdesolve(y*diff(U(x,y),x)+x*diff(U(x,y),y)=0, U(x,y) ) ')))

2 2
U(x, y) = _F1(-x + y )
8.16 Finite Difference Equations

MATLAB enables commands that allow you to solve finite difference equations and recurrence equations in general,
the prior use of the command maple is always required. We have:
rsolve (reqn, fnc (var))
Solves the recurrence equation reqn for the function fnc(var). You can use the
command asympt to find asymptotic solutions.
rsolve (expr, fnc (var))
Solves the recurrence equation expr = 0
rsolve ({reqn1,..., reqnm}, {fnc1 (var1),..., fncm (varm)})
Solves the set of recurrence equations given by the specified functions
rsolve ({reqn1,..., reqnn}, {fnc1,..., fncm}, ‘genfnc’ (var))
(for linear recurrences with constant coefficients and systems of linear
recurrences with constant coefficients) returns the generating functions fnc1,...,
fncn of the sequences defined by reqn1,…,reqnn.
rsolve({reqn1,...,reqnn},{fnc1,...,fncm},‘makeproc’)
(for any single recurrence that can be uniquely solved for the highest indexed
function) returns a procedure body for evaluating the function defined by
reqn1,...,reqnn.
REsol (reqn, fnc (var), {cond1,..., condm})
Represents the solution of the inert recurrence equation reqn for the function fnc
(var) subject to the specified initial conditions cond1,..., condn
(with (LREtools))
Loads into the memory a library of commands for work with linear recurrence
equations
constcoeffsol (reqn, fnc (var), {cond1,..., condm})
Returns the solution of the recurrence equation with constant coefficients reqn
according to fnc(var) and with initial conditions {cond1,..., condn}
209
www.pdfgrip.com
constcoeffsol(reqn,fnc(var),{cond1,...,condm},output=basis)
Returns a basis for the solutions of reqn
hypergeomsols(reqn,fnc(var),{cond1,...,condm})
Returns hypergeometric solutions of reqn. Also supports the option output =
basis. With the option output = onesol gives a simple solution.
polysols (reqn, fnc (var), {cond1,..., condm})
Returns polynomial solutions of reqn. Supports the additional options
output = basis and output = onesol.
ratpolysols (reqn, fnc (var), {cond1,..., condm})
Returns rational polynomial solutions of reqn with the given initial conditions
REcreate(reqn,fnc(var),{cond1,...,condn})
Creates a recurrence equation for reqn
REcreate ({reqn1,..., reqnn}, {(var) fnc1,..., fncn (var)}, {cond1,..., condm})
Creates n recurrence equations
REcontent (reqn, fnc (var), {cond1,..., condm})
Returns the content of the recurrence equation reqn
REprimpart(reqn,fnc(var),{cond1,...,condm})
Returns the primitive part of the recurrence equation
REreduceorder (reqn, fnc (var), {cond1,..., condm}, expr)
Reduction of order for the recurrent equation reqn where expr is a partial solution
REreduceorder (reqn, fnc (var), {cond1,..., condm}, [expr1,..., exprn])
Specifies n partial solutions
REtoDE (reqn, fnc (var), {cond1,..., condm})
Converts a recurrence equation reqn to a differential equation
REtodelta (reqn, fnc (var), {cond1,..., condm})
Converts the recurrence equation reqn to a difference equation
REtoproc(reqn,fnc(var),{cond1,...,condm})
Converts the recurrence equation reqn into a procedure
shift(expr, var, int)
Coincides with subs(var=var+int,expr)
shift(expr, var)
Takes int = 1 in the above by default
210
www.pdfgrip.com
Delta (expr, var, n)

Executes the difference of order n for the expression expr with respect to the
variable var
Delta (expr, var, n)
Executes the difference of order 1. We have that shift(expr,var,1)-expr = delta
(expr, var)

>> pretty(sym(maple('LREtools[REtodelta](u(n+2)-2*u(n+1)+u(n),u(n),{})')))

2
LREtools[Delta][n]

>> maple('rec:= a(n+2)-(2*n+1)*a(n+1)/n+n*a(n)/(n-1)=n*(n+1): ')
>> maple('re:= LREtools[REcreate](rec, a(n), { })')

ans =

re := RESol({a(n+2)*n^2-a(n+2)*n-2*a(n+1)*n^2+a(n+1)*n+a(n+1)+n^2*a(n) = n^4-n^2},{a(n)},{a(3)
= a(3), a(2) = a(2), a(1) = 0},INFO)

>> pretty(sym(maple('LREtools[polysols](re, output=basis) ')))

4 3 2
1/9 n - 5/18 n - 1/9 n + _C[1] n + 5/18 - _C[1]

>> pretty(sym(maple('LREtools[REtoDE]((18+12*n)*t(n)+(-20-7*n)*t(n+1)+ (4+n)*t(n+2),t(n), {t(0) = 0,
t(1) = 2}, y(z)) ')))

2 3 2
DESol({(-18 z + 13 z - 2) y + (-12 z + 7 z - z) D(y) + 6 z}, {y},

{y(0) = 0, D(y)(0) = 2})

>> maple('rec:= a(n+4) = -a(n+3)+3*a(n+2)+5*a(n+1)+2*a(n): ')
>> pretty(sym(maple('constcoeffsol(rec, a(n), { })')))

n
-2 (- 1/18 a(1) - 1/54 a(0) - 1/54 a(3) - 1/18 a(2)) 2

/ 68 13 \ n
+ |14/9 a(1) + -- a(0) - -- a(3) - 4/9 a(2)| (-1)
\ 27 27 /

n
+ (- 8/3 a(1) - 20/9 a(0) + 7/9 a(3) + 1/3 a(2)) (-1) (n + 1)

n
+ 1/2 (a(1) + 2/3 a(0) - 1/3 a(3)) (-1) (n + 1) (n + 2)

211
www.pdfgrip.com
Exercise 8-16
Find the solutions to the following recurrence equations:

ym + 1 = mym + (m + 1)! y1 = 2,
y2n = 4yn + 5 y1 = a,
yn+2 − 3yn+1 + 2yn = 4n y0 = y1 =1

>> pretty(sym(maple('rsolve({y(m+1)=m*y(m)+(m+1)!,y(1)=2},y)')))

2
1/2 GAMMA(m) (m + m + 2)

>> pretty(sym(maple('rsolve({y(2*n)=4*y(n)+5,y(1)=a},y) ')))

log(n)
------ + 1
2 2 log(2)
a n + n (- 20/3 (1/4) + 5/3)

>> pretty(sym(maple('rsolve({y(n+2)-3*y(n+1)+2*y(n)=4^n,y(0)=1,y(1)=1},y) ')))

n n
4/3 - 1/2 2 + 1/6 4

Exercise 8-17
Find the general term of the sequences of real numbers defined by the following recurrence laws:
xn −nxm* xn *xn + 1 = xn + 1 x0 = 1,
yn + 2 −2yn + 1 + 5yn = cos(3n) y0 = y1 = 1

>> pretty(sym(maple('rsolve({x(n)-n*x(n)*x(n+1)=x(n+1),x(0)=1},x) ')))

2
----------
2
n - n + 2

>> pretty(sym(maple('simplify(evalf(rsolve({x(n+2)-2*x(n+1)+5*x(n)=
=cos(3*n), x(0)=1,x(1)=1},x))) ')))

.4373424525 exp((.8047189562 - 1.107148718 I) n)

+ .4373424525 exp((.8047189562 + 1.107148718 I) n)

212
www.pdfgrip.com
3 3
- .9160727930 cos(n - 3.) + .4362567143 cos(n - 5.)

3 3
+ .1725076052 cos(n - 2.) - .1986350636 cos(n - 4.)

- .06265675740 I exp((.8047189562 + 1.107148718 I) n)

+ .06265675750 I exp((.8047189562 - 1.107148718 I) n)

+ .7931668115 cos(n - 3.) + .07520876265 cos(n - 2.)

213
www.pdfgrip.com
MATLAB Differential and

Integral Calculus
César Pérez López

www.pdfgrip.com
MATLAB Differential and Integral Calculus

Copyright © 2014 by César Pérez López
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material
is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval,
electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.
Exempted from this legal reservation are brief excerpts in connection with reviews or scholarly analysis or material
supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the
purchaser of the work. Duplication of this publication or parts thereof is permitted only under the provisions of the
Copyright Law of the Publisher’s location, in its current version, and permission for use must always be obtained from
Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations are
liable to prosecution under the respective Copyright Law.
ISBN-13 (pbk): 978-1-4842-0305-7
ISBN-13 (electronic): 978-1-4842-0304-0
Trademarked names, logos, and images may appear in this book. Rather than use a trademark symbol with every
occurrence of a trademarked name, logo, or image we use the names, logos, and images only in an editorial fashion
and to the benefit of the trademark owner, with no intention of infringement of the trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified
as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.
While the advice and information in this book are believed to be true and accurate at the date of publication, neither
the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may
be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.
Publisher: Heinz Weinheimer
Lead Editor: Dominic Shakeshaft
Editorial Board: Steve Anglin, Mark Beckner, Ewan Buckingham, Gary Cornell, Louise Corrigan,
Jim DeWolf, Jonathan Gennick, Robert Hutchinson, Michelle Lowman, James Markham, Matthew Moodie,
Jeff Olson, Jeffrey Pepper, Douglas Pundick, Ben Renow-Clarke, Dominic Shakeshaft, Gwenan Spearing,
Matt Wade, Steve Weiss
Coordinating Editor: Jill Balzano
Copy Editor: Barnaby Sheppard
Compositor: SPi Global
Indexer: SPi Global
Artist: SPi Global
Cover Designer: Anna Ishchenko
Distributed to the book trade worldwide by Springer Science+Business Media New York, 233 Spring Street, 6th Floor,
New York, NY 10013. Phone 1-800-SPRINGER, fax (201) 348-4505, e-mail orders-ny@springer-sbm.com, or
visit www.springeronline.com. Apress Media, LLC is a California LLC and the sole member (owner) is Springer
Science + Business Media Finance Inc (SSBM Finance Inc). SSBM Finance Inc is a Delaware corporation.
For information on translations, please e-mail rights@apress.com, or visit www.apress.com.
Apress and friends of ED books may be purchased in bulk for academic, corporate, or promotional use. eBook
versions and licenses are also available for most titles. For more information, reference our Special Bulk Sales–eBook
Licensing web page at www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this text is available to readers at
www.apress.com. For detailed information about how to locate your book’s source code, go to
www.apress.com/source-code/.
www.pdfgrip.com
Contents
About the Author��ix
■■Chapter 1: Introduction and the MATLAB Environment��1

1.1 Numerical Computation with MATLAB��1
1.2 Symbolic Computation with MATLAB��2
1.3 MATLAB and Maple��4
1.4 Graphics with MATLAB��5
1.5 Environment and General Notation��10
1.6 Help with Commands��11
1.7 Commands to Exit and Escape to the MS-DOS Environment��14
1.8 MATLAB and Programming��15
■■Chapter 2: Limits and Continuity. One and Several Variables��17
2.1 Limits of Sequences��18
2.2 Limits of Functions. Lateral Limits��20
2.3 Sequences of Functions��23
2.4 Continuity��28
2.5 Several Variables: Limits and Continuity. Characterization Theorems��33
2.6 Iterated and Directional Limits��35
2.7 Continuity in Several Variables��40
v
www.pdfgrip.com
■ Contents
■■Chapter 3: Numerical Series and Power Series��45

3.1 Series. Convergence Criteria��45
3.2 Numerical Series with Non-Negative Terms��45
3.3 Alternating Numerical Series��55
3.4 Power Series��56
3.5 Formal Power Series��58
3.6 Power Series Expansions and Functions��61
3.7 Taylor, Laurent, Padé and Chebyshev Expansions��63
■■Chapter 4: Derivatives and Applications. One and Several Variables��69
4.1 The Concept of the Derivative��69
4.2 Calculating Derivatives��72
4.3 Tangents, Asymptotes, Concavity, Convexity, Maxima and Minima,
Inflection Points and Growth��75
4.4 Applications to Practical Problems��82
4.5 Partial Derivatives��84
4.6 Implicit Differentiation��85
4.7 Differentiation of Functions of Several Variables��89
4.8 Maxima and Minima of Functions of Several Variables��98
4.9 Conditional Minima and Maxima. The Method of “Lagrange Multipliers”��102
4.10 Some Applications of Maxima and Minima in Several Variables��106
■■Chapter 5: Vector Differential Calculus and Theorems in Several Variables��109
5.1 Concepts of Vector Differential Calculus��109
5.2 The Chain Rule��110
5.3 The Implicit Function Theorem��113
5.4 The Inverse Function Theorem��115
5.5 The Change of Variables Theorem��118
5.6 Taylor’s Theorem with n Variables��119
5.7 Vector Fields. Curl, Divergence and the Laplacian��120
5.8 Coordinate Transformation��123
vi
www.pdfgrip.com
■ Contents
■■Chapter 6: Integration and Applications��129

6.1 Indefinite Integrals��132
6.2 Integration by Substitution (or Change of Variables)��133
6.3 Integration by Parts��135
6.4 Integration by Reduction and Cyclic Integration��137
6.5 Definite Integrals��138
6.6 Curve Arc Length��138
6.7 The Area Enclosed between Curves��141
6.8 Surfaces of Revolution��149
6.9 Volumes of Revolution��151
6.10 Curvilinear Integrals��154
6.11 Approximate Numerical Integration��155
6.12 Improper Integrals��156
6.13 Parameter Dependent Integrals��157
6.14 The Riemann Integral��159
■■Chapter 7: Integration in Several Variables and Applications��163
7.1 Areas and Double Integrals��165
7.2 Surface Area by Double Integration��169
7.3 Volume Calculation by Double Integrals��171
7.4 Volume Calculation and Triple Integrals��174
7.5 Green’s Theorem��176
7.6 The Divergence Theorem��178
7.7 Stokes’ Theorem��178
■■Chapter 8: Differential Equations��181
8.1 Separation of Variables��184
8.2 Homogeneous Differential Equations��186
8.3 Exact Differential Equations��188
8.4 Linear Differential Equations��191
vii
www.pdfgrip.com
■ Contents
8.5 Ordinary High-Order Equations��191

8.6 Higher-Order Linear Homogeneous Equations with Constant Coefficients��193
8.7 Non-Homogeneous Equations with Constant Coefficients. Variation of Parameters��195
8.8 Non-Homogeneous Linear Equations with Variable Coefficients.
Cauchy-Euler Equations��199
8.9 The Laplace Transform��200
8.10 Systems of Linear Homogeneous Equations with Constant Coefficients��202
8.11 Systems of Linear Non-Homogeneous Equations with Constant Coefficients��203
8.12 Higher Order Equations and Approximation Methods��204
8.13 The Taylor Series Method��204
8.14 The Runge–Kutta Method��206
8.15 Partial Differential Equations��208
8.16 Finite Difference Equations��209
viii
www.pdfgrip.com
About the Author
César Pérez López is a Professor at the Department of Statistics and Operations Research at the University of
Madrid. César is also a Mathematician and Economist at the National Statistics Institute (INE) in Madrid, a body
which belongs to the Superior Systems and Information Technology Department of the Spanish Government.
César also currently works at the Institute for Fiscal Studies in Madrid.
ix
www.pdfgrip.com
Coming Soon
• MATLAB Programming for Numerical Analysis, 978-1-4842-0296-8

• MATLAB Differential Equations, 978-1-4842-0311-8
• MATLAB Control Systems Engineering, 978-1-4842-0290-6
• MATLAB Linear Algebra, 978-1-4842-0323-1
• MATLAB Matrix Algebra, 978-1-4842-0308-8
xi

MATLAB Differential and Integral Calculus

Uploaded by

Copyright:

Available Formats

MATLAB Differential and Integral Calculus

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MATLAB Differential and Integral Calculus

Uploaded by

Copyright:

Available Formats

www.pdfgrip.

For your convenience Apress has placed some of the front

■■Chapter 1: Introduction and the MATLAB Environment�����������������������������������������������������1

Introduction and the MATLAB

1.1 Numerical Computation with MATLAB

1.2 Symbolic Computation with MATLAB

1.3 MATLAB and Maple

To use a Maple command from Matlab, the syntax is as follows:

1.4 Graphics with MATLAB

3. We can generate a graph of the surface defined by the function z = sin(sqrt(x^2+y^2)) /

We will see more on these concepts later.

1.5 Environment and General Notation

1.6 Help with Commands

matlab\datafun - Data analysis and Fourier transforms.

atan - Inverse tangent.

abs - Absolute value.

>> lookfor complex

1.7 Commands to Exit and Escape to the MS-DOS Environment

1.8 MATLAB and Programming

Limits and Continuity.

2.1 Limits of Sequences

Calculate the following limits:

Calculate the following limits:

The first two examples are indeterminate of the form 1∞:

2.2 Limits of Functions. Lateral Limits

Calculate the following limits:

Calculate the following function limits:

The first limit is calculated as follows:

2.3 Sequences of Functions

Calculate the limit function of the following sequences of functions:

Calculate the limit of the following sequences of functions:

>> pretty(sym(maple('limit(piecewise(x>=-n and x<=n,1,0),n=infinity)')))

MATLAB provides the following commands relating to continuity:

Study the continuity of the following functions of a real variable:

Study the continuity of the following function of a real variable:

Study the continuity of the following function of a real variable:

2.5 Several Variables: Limits and Continuity. Characterization

lim ( a1n ) = a1 , lim ( a2 n ) = a2 ,..., lim ( amn ) = am .

Calculate the limit of the following three-dimensional sequence:

Calculate the limit as n→∞ of the following four-dimensional sequence:

For the function f : R ® R 2 defined below, find lim f ( x ) :

For the function f : R 2 ® R 2 defined below, find lim f (x , y) :

é 2(1 - cos( y )) sin( x ) x tan( y ) ù

[inf, exp (1) - 1]

2.6 Iterated and Directional Limits

Find lim f ( x , y ) for the function f : R 2 → R defined by:

Find lim f ( x , y ) for the function f :R 2 → R defined by:

Find lim f ( x , y ) for the function f :R 2 → R defined by:

Find lim f ( x , y ) for the function f :R 2 → R defined by:

Find lim f ( x , y ) for the function f : R 2 → R defined by:

2.7 Continuity in Several Variables

lim f (a1 , a2 ,¼, an ).

Let the function f : R 2→ R be defined by:

Find lim f ( x , y ) and study the continuity of f.

Define the function f : R 2 ® R by

and study its continuity at the point (1,2).

Consider the function f : R 2 ® R defined by:

Study the continuity of f at the point (0,0).

Numerical Series and

3.1 Series. Convergence Criteria

■■Chapter 1: Introduction and the MATLAB Environment��1