__________________________________________________________________________

MATLAB FOR COMPUTATIONAL PHYSICS

R J BALLAGH, PHYSICS DEPARTMENT, UNIVERSITY OF OTAGO
__________________________________________________________________________

INTRODUCTION

Matlab is an interactive software system for doing scientific calculations. It has many built in
mathematical functions and plotting capabilities that allow you easily to carry out computations
that arise in Physics. It can be used rather like a calculator, or for more complex problems, it can
be used to write programmes.

The purpose of these notes is to help you begin to use MATLAB to solve some physics
problems. They can best be used hands-on. You should work at the computer as you read the
notes and freely experiment with examples.

At the end of the course, you should have developed enough skills to allow you to use MATLAB
on your own, to tackle problems that would be too difficult or tedious to do "by hand".

1
MATLAB: KEY IDEAS

COMMANDS
To do work in MATLAB, you type commands at the >> prompt, followed by the return key.
Often these commands will look like standard arithmetic, or function calls similar to many other
computer languages.

Later, you will learn how to store a sequence of commands in a file, and to execute them
automatically, just as if they had been entered from the keyboard. Such files are called script
files in MATLAB. It is also possible to extend the commands that MATLAB knows about by
writing function files. For example you might write a function file that calculates the volume of
a sphere, given the value of the radius.

THE BASIC OBJECTS: MATRICES AND VARIABLES

Matlab works with essentially only one kind of object, a matrix (hence its name; MATrix
LABoratory). A matrix is simply a rectangular collection of real or complex numbers, and is
said to be of size m n if there are m rows and n columns. In this course we will find that many
physical problems of interest can be expressed in terms of matrices.

In some situations, MATLAB attaches special meaning to 1-by-1 matrices, which are scalars,
and to matrices with only one row or one column, which are vectors.

SCALARS
Let's start with the simplest example, a scalar, which we'll call x (and which might represent the
distance of a ball from the ground). MATLAB doesn't know about x until we tell it. We can do
this by assigning a value to x, by typing at the prompt

x = 5

Note the response. x is called a variable, and we can store any numerical value in it.

Conventional decimal notation, with optional decimal point and leading minus sign, is used for
numbers. A power-of-ten scale factor can be included as a suffix. Here are some examples of
legal numbers:

3
9.6397238

-99
0.0001

1.60210e-20 (which represents 1.60210 1020)

6.0225e23 (which represents 6.0225 1023)

2
At the simplest level you can use MATLAB just like a calculator, using the following arithmetic
operators.

+ addition
subtraction
* multiplication
^ raise to a power
/ division

Exercise: Try x = 5*4, and some other examples that occur to you.

There are also a number of built in functions available to you in MATLAB, such as square root.
You can see how this works by typing

sqrt (x)

A complete list of built in functions is given in the manual, and later in these notes.

STATEMENTS, EXPRESSIONS AND VARIABLES

Now that you have broken the ice with the basic ideas, it is worthwhile to describe the structure
of MATLAB a little more formally.

MATLAB is an expression language; the expressions you type are interpreted and evaluated
once you press the return key. MATLAB statements are usually of the form

variable = expression

or simply

expression

Expressions are usually composed from operators, functions, and variable names. Evaluation of
the expression produces a matrix, which is then displayed on the screen and assigned to the
variable for future use. If the variable name and = sign are omitted, a variable ans (for answer)
is automatically created to which the result is assigned.

A statement is normally terminated with the return key. However, a statement can be continued
to the next line with three or more full stops followed by the return key. On the other hand,
several statements can be placed on a single line if separated by semicolons.

If the last character of a statement is a semicolon, you don't see the result, but the assignment is
carried out. This is essential in suppressing unwanted printing of intermediate results.

Exercise: Try some examples.

3
EVALUATING FORMULAS
You probably remember from elementary kinematics the equation for the distance y that an
object travels under constant acceleration a in a time t, starting with an initial velocity v0

1 2
y = v0 t + at (1)
2

This equation can be entered in MATLAB by typing

y = v0*t + (1/2)*a*t^2 (2)

I have used v0 to represent v0, but you may choose any name that
begins with an alphabetical character, and can include digits or
underscores up to a total of 19 characters.

Try it. What did you get? The problem is, that MATLAB doesn't yet know about the variables
v0,a t; you must first assign some numerical values to them. Do this choosing some values
that appeal to you. Now you can re-enter equation (2). (You can save yourself a lot of re-typing
by using the up arrow to recall a previous line to the command line, and then edit it as required.)

Now MATLAB returns the value of y. Is it correct? Did you make a good choice for the
assigned values of the variables v0, a and t, in order to allow you to reliably check that you
had entered the formula correctly?

Suppose we want to evaluate the distance y for a sequence of times, say t = 0,1,2,3,4. We could
repeat the above procedure for each value of t, and it's probably worth doing this once. Write
down your answers for later reference. But now you are about to glimpse some of the real power
of MATLAB. The sequence of times can be written as a matrix, which we can visualise as

t = (0 1 2 3 4) (3)

which has one row and five columns. All variables in MATLAB actually represent matrices, and
the easiest way to assign a matrix of values to the variable (for a small matrix such as above) is
by an explicit list of elements separated by blanks or commas, and surrounded by the square
brackets [ and ] . So to enter the matrix from line (3) into MATLAB we type

t = [0 1 2 3 4]

If there is more than one row, you can indicate the end of each row by a semicolon ; or the
return key. For example the matrix

F1 2 3 I
A = G4 6J
GH 7 5
8
J
9K
(4)

is entered by either of the statements

A = [1 2 3; 4 5 6; 7 8 9]

4
or
A = [
1 2 3
4 5 6
7 8 9 ]

Try it. Notice that if you used the lower case a instead of A you will have overwritten your
earlier assignment to the variable representing acceleration in equation (2).

MATLAB is case sensitive and distinguishes between

lower case and upper case.

If you haven't already entered the matrix of line (3) into your variable t, do so now. Notice this
is a one row matrix, which we sometimes call a vector in MATLAB. It is possible to get
MATLAB to calculate the sequence of y values corresponding to the sequence of t values using a
very similar expression to line (2). But first it may be helpful to recall a few important ideas
about matrices.

MATRIX CONCEPTS
A matrix is a two dimensional array of numbers (or symbols), arranged in a rectangular
form with (say) m rows and n columns. If m = n, we call the matrix square. Two
matrices, B and C, can be multiplied together in the order

BC

by the rule we can colloquially describe as "diving the rows of B into columns of C".
You should remember this from your algebra courses. However a necessary requirement
is that the number of columns of B is equal to the number of rows of C.

We can denote the element from the ith row and jth column of matrix A by using
subscripts Aij or parentheses, A(i,j). In line (4) for example A23 = 6.

If you are rusty on basic matrix concepts, spend some time reviewing them before your next
session.

MATRIX OPERATIONS
MATLAB is set up to do matrix operations, so if you have assigned values (or elements) to the
variables B and C, the expression

B * C

multiplies them according to the rule discussed above. The operation is defined whenever the
"inner dimensions" of the two matrices are the same.

Exercise: Explore some examples.

5
A scalar can also multiply, or be multiplied by a matrix e.g. Try

5 * A

Addition and subtraction of matrices are denoted by + and and are simply element-by-element
addition or subtraction of the corresponding entries. The operations are defined whenever the
matrices have the same dimensions.

Exercise: Enter a matrix D, where

F 01. 0.2 0.3 I

D = G 0.4 0.6 J
GH 0.7 0.5
0.8 0.9K
J
and then enter

F = A + D

Do you understand the result you get?

Addition and subtraction are also defined if one of the operands is a scalar, that is a 1-by-1
matrix. In this case, the scalar is added to or subtracted from all the elements of the other
operand. For example try

z = 1 + D

There are other matrix operations; matrix "division", matrix powers, that you will discover as
you explore MATLAB.

ARRAY OPERATIONS (ELEMENT-WISE OPERATIONS)

We will often want to perform an operation on each element of a vector while doing a
computation. Addition (+) and subtraction () already work this way. For example, if
w = [1 2 3] and u = [5 6 2], then w + u will return [6 8 5]. Matrix
multiplication is different; rows "dive into" columns; similarly matrix division and matrix
power operations have specifically defined linear algebraic meaning. But suppose we wanted to
multiply the first element of w by the first element of u, and the second element of w by the
second element of u, and so on. (We will shortly give a physical example where this is done).
We achieve this in MATLAB by preceding the * operator by the fullstop ., as in

w.*u

Exercise: Try it, as well as w*u

Similarly / or ^ can be preceded by a fullstop to specify elementwise division or power

operations. In MATLAB, these are called "array operations".

Exercise: Explore the array operation using w and u.

There is one subtlety you should be careful with. It is of course possible to take an element by
element power using a scalar base and matrix exponent. Let's put s = 2 and type

6
 s.^w

Make sure you understand the result that is returned. We could have done this more directly by
typing

2 .^w

but now the space between the digit 2 and the fullstop . is important. If it was not there, the
fullstop would be interpreted as a decimal point associated with the 2.

MATLAB would then see only the isolated ^ and would attempt to calculate a matrix power,
which in this case would result in an error message because the exponent matrix is non-square.
An alternative to the space is to use parentheses, forcing MATLAB to interpret the base as a
number before it carries out the array operation, (2).^w

Now let's return to our problem of an object travelling under constant acceleration, where we
were wishing to evaluate the distance y for a sequence of times. First type t to be sure that
your variable t is the vector given in equation (3).

Now re-enter equation (2) (or use the clipboard) but with t2 defined as an array operation

y = v0*t + (1/2)*a*t.^2

Note that because v0 and a are scalars, we didn't need to use a fullstop with their
multiplications (although you could do so try it).

PLOTTING RESULTS
A very powerful technique in physics is to graphically display results so they can be visualised,
and MATLAB has very convenient plotting capabilities. We can make a graph of y versus t
by the command

plot(t,y)

Do you understand the graph? It is simple to redo the graph with a different value of one of the
parameters. Change the value of the acceleration, re-evaluate the expression for y, and plot the
new results. Now change v0. Then try a different sequence of times.

It is easy also to add a title and axes label to your graph, by entering (after the plot command)

title ('distance versus time')

xlabel ('time')
ylabel ('distance')

Notice that MATLAB refers to the vertical axis as y, and the horizontal axis as x, in the normal
way. The single quotes around the characters you are putting on the graph are necessary to
indicate they are to be treated as just a string of characters (and not variables, etc)

7
MORE MATRIX STUFF
So far we have dealt mainly with one-dimensional matrices but it is possible to use two-
dimensional matrices, and calculate y distances for a sequence of times, and a range of
accelerations, in one calculation. To do this we will need to know about the matrix transpose
operation.

MATRIX TRANSPOSE
Formally, the transpose of a matrix A, which we write A', is another matrix with elements
defined as

A'ij = Aji

You can think of this as "rotating" A to make a new matrix A' whose columns are the rows of A.
MATLAB does this, naturally enough, with the apostrophe symbol. Type

to recall the matrix you entered earlier in the session (see Eq.(4)). Now type

A'

Exercise: Try it also with the row vectors you have previously entered. What is the
transpose of a column vector?

The row vectors (or 1 3 matrices) w = [ 1 2 3 ], u = [ 5 6 2 ] that you entered

earlier can not be matrix multiplied. Try w*u. Why does it fail? However w*u' gives an
answer. This in fact is the inner product of the two vectors. Do you understand why? Now try
u'*w. Do you see what is happening? Let's return to our kinematic problem, and calculate the y
distances for the sequence of times t = (0,, 1, 2, 3) seconds for the cases where the acceleration a
= 2 m/s2 and also a = 4 m/s2. Let's begin by assuming v0 = 0, and using y = 1 at 2 . Just as we
2
represented the sequence of times as a vector, stored in the variable t, we can represent the
range of accelerations as a vector stored in the variable a

a = [2 4]

In scalar arithmetic, the order of factors in a multiplication is unimportant. In matrix

multiplication the order is crucial. Typing

y = 0.5*(t.^2)'*a

produces a 4 2 matrix whose first column gives the y distances for a = 2 and the second column
gives the y distances for a = 4. Make sure you understand why (you don't really need the
brackets, but they make it clear what you are doing). Now include the contribution of the initial
velocity. any success? You need the v0t part to also be a 4 2 matrix, where column one applies
for a = 2 and column 2 for a = 4. There are many ways to do this, and here is one

y = v0*t'*[1 1] + 0.5*t.^2'*a

By the way, it is worthwhile to know that in MATLAB you can easily make up an m n matrix
with every element equal 1 by the command ones(m,n). Other simple matrices (e.g. the
identity), are equally simple to make; see the reference manual.

8
You can plot the distances for each acceleration together on one plot by typing

plot(t,y)

which causes each column of y to be plotted successively against the vector t.

Exercise: Repeat the distance calculations above, but this time producing the y distances
in rows rather than columns. Plot them out.