An Introduction to Matlab: Part 3

This lecture assumes that you have already worked through part 1 and part 2. You should be able to use many basic Matlab
commands and use Matlab as a calculator on scalar variables. You should be able to create and execute a script le. This
lecture covers
Creating row/column vectors
Vector operations
Applying functions to vectors
Component wise arithmetic
Creating vectors

This section goes through creating row and column vectors by typing each element, by using patterns, and by using a few
built in functions.
1.

Typing in vectors explicitly:

Here we learn how to type vectors.

(a) Open Matlab. If you already have it open, type clear all; in the Command Window. You may do everything here
in a script le or on the Command Window. I would suggest simply using the Command Window for now so you
don't have to rerun you le everytime we make a change.
(b) Let us create a row vector consisting of the elements 1, 3, 5 and 7. To do this, we use the bracket notation []. We
assign this vector to the variable u.
u=[1,3,5,7]

Instead of using commas, try simply inserting a space between each element. Type
u=[1 3 5 7]

Note that they are the same thing. Also note that this is a row vector (Matlab prints out the numbers in a row).
(c) Let us create a column vector consisting of the same elements 1, 3, 5 and 7. For a row vector, we put either spaces
or commas in between elements. For a column vector, we put a semicolon between each element. Note that a
semicolon is also used to supress output, but Matlab is smart enough to know that a semicolon within brackets []
means to change rows. Type
v=[1;3;5;7]

Note that Matlab displays the elements of v as a column.

Typing in vectors using a pattern:

We create vectors which are large and correspond to a distinct pattern.

(a) Suppose we want to create a row vector, u, that contains elements 1 through 100 (ie u = [1, 2, . . . , 99, 100]). How
can we do this? Well, you can type in each element, but that seems awfully silly. Matlab has a nice way to do
this. Type
u=[1:1:100]

This notation means u goes from 1 by 1 to 100. Create the vector u = [1, 3, 5, 7, 9, 11] using this notation. How
do you think you would create the vector u = [15, 12, 9, 6, 3, 0]? Well, it goes from 15 by 3 to 0. So we type
u=[15:-3:0]

We can also use negative numbers. Create a vector that goes from 15 to 0 by 3.
(b) Note that using the notation above (u=[15:-3:0] ) created a row vector. Let us say we want a column vector
instead. The easiest way to do this is to use the transpose operator, which is the single quote: '. Type:
u=[15:-3:0]
v=u'

What happened? Well, u is created as a row vector, then v is the transpose of u, which is a column vector. We
could have simply done this on one line also by typing
v = [15:-3:0]'

Try to create a column vector that goes from 2 to 16 by 2.

3.

Creating a zero vector, one vector, or random vector:

of completely random entries (between 0 and 1)

Here we discuss how to create a vector of all zeros, all ones, or

(a) To create a vector of all zeros, we use the zeros function. It takes two (for now) arguments. These represent the
size of the vector you wish to create. Type:
u = zeros(5,1)

You get a column vector with 5 zeros. Try to create a row vector with 10 entries.
(b) To create a vector of all ones, we use the ones function. It has the same structure as the zeros function:
v = ones(1,6)

(c) You can also create a vector with random entries (these random entries are randomly drawn from between 0 and
1) using the rand function:
u = rand(4,1)

Vector operations

This section goes through basic vector arithmetic and how to create new vectors using parts of old vectors.
1.

Vector addition and scalar multiplication:

(a) Type clear

all;

We learn how to add vectors and scale vectors.

Create three row vectors, u, v and w by doing the following:

u = [1 3 5 7]
v = [0 1 2 4]
w = [2 4 6 8 10]

Add vectors u and v by typing:

Note that your result is another vector. We could assign it a varible by typing r=u+v .
Now add v and w. What happens? We get an error because they are not the same size. Recall that we can only
add vectors of the same size.
Create a new vector, v1, that is the transpose of v , again by typing :
v1=v';

Try adding v1 and v. What happens? An error again. Why? Because row vectors and column vectors cannot be
added together.
(b) How do you suppose we multiply a vector by a scalar? Well, exactly how you would expect! Let us multiply w by
2 by typing:
2*w

Find u and e2 v by doing the same.

New vectors from old vectors:

extract part of an old vector.

Here we'll learn how to incorporate an old vector as part of a new vector, or how to

(a) We'll use the same u, v, and w from above, so recreate the vectors if need be. Let us say that I only want to look
at the 2nd element (component) of vector v . How might I do this? Well, vectors are indexed from 1 until their
size (4 in this case) and are accessed using parenthesis. So typing:
v(2)

returns the 2nd element of v . Use this idea to nd the 1st and 4th components of v .
(b) Let us say that I want more than one element of v . Say I want to get a new vector that is the same as v , but without
the 2nd element. So I want a new vector, call it v1, that is size 3 an contains the 1st , 3rd and 4th components of
v . I could type:
v1=[v(1) v(3) v(4)]

but for very large vectors, this will not be feasible. Another way I could do this is by typing:
2

v1=v([1 3 4])

which says take only the parts with index in [1 3 4]. This is still a bit dicult for large vectors. Luckily, we can
use the colon notation (like when we dened u=[15:-3:0]). So I can instead type:
v1=v([1 3:4])

Create a new vector r that goes from 3 by 3 to 99. From this vector, create another vector r1 that contains the
rst 3 elements of r, the 15th element, the 18th element and the last 4 elements (elements 30 through 33).
(c) We could also take elements out of order. For example, using the r dened previously, lets say we wanted a new
vector r1 that was the last element (33rd ), the 10th element, then elements 5, 4, 3 and 2. You would do this exactly
as expected:
v1=v([33 10 5:-1:2]).

(d) Now, let us consider the problem of enlarging a vector, rather than taking part of a previous vector. Let us create
the vector t = [0, 1, 2, 4, 6, 8] by taking v = [0, 1, 2, 4] and adding 6 and 8 to the end. There are two ways to do
this:
i. Simply dene t = v then t(5) = 6 and t(6) = 8. This takes previously undened value of t and denes them,
thus enlarging t. Try this.
ii. Dene a new vector that is exactly what you want, v , with a 6 and 8 tacked on:
t=[v,6,8]

Use this idea to create a new vector d = [2, 0, 1, 2, 4, 10] by adding 2 to the beginning of v and 10 to the end.
You can also combine two vectors, by doing something like:
s=[v,t]

(e) Note that we have used row vectors, but all of the above operations hold for column vectors as well. The dierence
is that notation such as
s=[v,t]

will not make sense if v and t are arbitrary column vectors. Instead we'd want to use
s=[v ; t]

to make sure the rows from t appear after the rows from v .
Applying functions to vectors

As most of you have seen in your classes, there are many useful functions that apply to vectors. For example, one might
want to know the norm of a vector or the dimension of a vector. In this section, we use a handful of these functions. These
are by no means all the useful functions, but just a small subset.
1.

Functions on a single vector:

vectors.

Here we discuss functions that take a single vector as the argument, or apply to single

(a) Type clear all; and dene a = [1, 0, 1], b = [0, 0, 1], c = [1 : 1 : 10] and d = [2 : 2 : 20].
(b) Recall the norm of a vector x = [x1 , . . . , xn ] is dened as
kxk =

x21 + + x2n .

The Matlab function for the norm is, oddly enough, norm. To nd the norm of c, type:
norm(c)

Using norm, we can create a unit vector simply by dividing a given vector by its norm. Create a vector v that is
a unit vector in the same direction as d.
(c) To nd the length or size (dimension) of a vector, we can use the functions length or size. length returns one
argument, while size returns two. Type:
[n m ] = size(c)
k = length(c)

Notice the dierence between the two functions.

2.

Functions on multiple vectors:

Here we discuss functions that take multiple vectors as arguments

(a) We will be using the same vectors as previously, so redene a, b, c and d if necessary.
(b) Let us start with the dot product. Recall for two vectors of the same size, c d is the sum of the product of each
individual component. As you may guess, the dot product function is dot. Try:
dot(c,d)

What about dot(a,c)? What happens?

What happens if you try cross(c,d)?

help FunctionName

for more

Find minimum/maximum: min(c), max(c)

Component wise arithmetic

We can also manipulate vectors using component wise arithmetic . The reason this is in bold print and gets its own section
is because this is a very hard concept for beginning Matlab users, but one that is essential to fully understand Matlab. Let
us say that I want to square the entries of the vector d. How can I do this? Well, try:
d^2

What happens? It doesn't work, why not? Because I cannot multiply d times itself, because that type of vector multiplication
makes no sense. Matlab has a way to take care of this. We add a period . before the operation. Try:
d.^2

What happened? Matlab applied the operation ^2 to each part of d. That's what the . meant.
Using the same idea, we can multiply each entry of d by the corresponding entry of c using
d.*c

Again, this means multiply

each component

of d to the

corresponding component

Create a new vector that contains the entries of c divided by the entries of d.

of c. Note that this only makes sense if c