Scientific Computing Lecture Series

Introduction to MATLAB Programming

Hamdullah Yücel*
∗
Scientific Computing, Institute of Applied Mathematics

Lecture I
Basic Commands and Syntax, Arrays and Matrices

H. Yücel (METU) MATLAB Lecture I 1 / 46

Computational Science

Experiment

Theory Computational
Science

Computational Science now constitutes what many call the third pillar of the scientific
enterprise, a peer alongside theory and physical experimentation.

Report to the President:”Computational Science : Ensuring America’s Competitiveness”,

June 2005.
H. Yücel (METU) MATLAB Lecture I 2 / 46
Scientific Computing

Mathematics

SC
Computer
Science Natural
Sciences

Scientific Computing

= Computational Science

= Computational Science and Engineering

= Scientific Computation

= Computational Mathematics

0 Quoted by https://en.wikipedia.org/wiki/Computational science

H. Yücel (METU) MATLAB Lecture I 3 / 46
Scientific Computing Program

Computational Computational
Biology Mechanics

Computational Mathematical Computational

Model Order
Fluid Modelling and Electro-
Reduction
Dynamics Application magnetic

Uncertainty Adaptive
Quan- Mesh
Portfolio
tification Generation
Optimization Algorithmic
Computational
Trading
Mathematics
& Simulation SCIENTIFIC
PDE-
Constrained
COMPUTING Computational
Optimization
Finance
High
(Non)Linear
Performance
Programming
Computing

Big Data

Data Science

Operations
Research

Statistical
Learning and
Data Mining

H. Yücel (METU) MATLAB Lecture I 4 / 46

For More Information

http://iam.metu.edu.tr/scientific-computing
https://www.facebook.com/SCiamMETU/

H. Yücel (METU) MATLAB Lecture I 5 / 46

Lecture Information

MATLAB Lecture Series is organized by members of Scientific Computing Program of IAM:

February 25: Hamdullah Yücel

Basic Commands and Syntax, Arrays and Matrices

February 26: M. Alp Üreten

Scripts and Functions, Control Loops and Advanced Data Structures

February 28: Süleyman Yıldız

Graphics, Visualizations and Symbolic Toolbox

H. Yücel (METU) MATLAB Lecture I 6 / 46

Lecture I–Outline

1 Introduction to MATLAB

2 Data Structures: Arrays and Matrices

3 Operators

4 Sparse Matrices

H. Yücel (METU) MATLAB Lecture I 7 / 46

1 Introduction to MATLAB

2 Data Structures: Arrays and Matrices

3 Operators

4 Sparse Matrices

H. Yücel (METU) MATLAB Lecture I 8 / 46

What is MATLAB ?

Matlab is a high–level language and interactive environment that enables you

to perform computationally intensive tasks. It was originally designed for
solving linear algebra type problems using matrices. It’s name is derived from
MATrix LABoratory.

H. Yücel (METU) MATLAB Lecture I 9 / 46

MATLAB System

Desktop Tools and Development Environment

Includes the MATLAB desktop and Command Window, an editor and
debugger, a code analyzer, browsers for viewing help, the workspace, files, and
other tools.
Mathematical Function Library
Vast collection of computational algorithms ranging from elementary
functions, like sine, cosine, and complex arithmetic, to more sophisticated
functions like matrix inverse, matrix eigenvalues, Bessel functions, and fast
Fourier transforms.
Language
The MATLAB language is a high-level matrix/array language with control flow
statements, functions, data structures, input/output, and object-oriented
programming features.
H. Yücel (METU) MATLAB Lecture I 10 / 46
Continue...

Graphics
MATLAB has extensive facilities for displaying vectors and matrices as graphs,
as well as editing and printing these graphs. It also includes functions that
allow you to customize the appearance of graphics as well as build complete
graphical user interfaces on your MATLAB applications.

External Interfaces
The external interfaces library allows you to write C and Fortran programs that
interact with MATLAB.

H. Yücel (METU) MATLAB Lecture I 11 / 46

MATLAB Interface

Command Window: Here you can give MATLAB commands typed at the prompt, >>.

Current Directory: Directory where MATLAB looks for files.

Workspace: Shows what variable names are currently defined and some info about their
contents.

Command History: History of your commands.

H. Yücel (METU) MATLAB Lecture I 12 / 46
Helps/Docs

MATLAB is huge! - there is no way to remember everything you will need to

know.
help command - shows in the Commmand Window all the ways in which you
can use the command.
doc command - brings up more extensive help in a separate window.
lookfor command- searches for the keyword.
>> help sin
SIN Sine of argument in radians.
SIN(X) is the sine of the elements of X.
See also asin, sind.
Overloaded methods:
codistributed/sin
Reference page in Help browser
doc sin

H. Yücel (METU) MATLAB Lecture I 13 / 46

Basic Commands

MATLAB records in the workspace and command history everything you

write in the command window, so:
clear variable
deletes variable from memory (and workspace)
clear all
deletes all variables from memory (and workspace)
clc
cleans command window
save
save variables to a file (.mat format)
load
load variable bindings into the environment (look at workspace, the variables a
is back)

MATLAB’s command window works like a Linux terminal

cd, mkdir, rmdir, ls, . . .
H. Yücel (METU) MATLAB Lecture I 14 / 46
Basic Commands

Some commands used to interact with MATLAB

what
returns the MATLAB files (.m , .mat) in the current directory

who
returns the variables in your workspace

whos
returns the variables in the workspace with additional info (size, dimensions)

Try typing why in the command window. You will see that MATLAB is also a
Philosopher!

H. Yücel (METU) MATLAB Lecture I 15 / 46

1 Introduction to MATLAB

2 Data Structures: Arrays and Matrices

3 Operators

4 Sparse Matrices

H. Yücel (METU) MATLAB Lecture I 16 / 46

Variables

MATLAB is a weakly typed language

No need to initialize variables!
Just assign some value to a variable name, and MATLAB will automagically
understand its type
x =3 double
x = ’hello’ char
MATLAB supports various types, the most often used are
64-bit double (default)
16-bit char

Most variables you will deal with will be vectors or matrices of doubles or
chars
Other types are also supported: complex, symbolic, 16-bit and 8-bit integers,
etc.
H. Yücel (METU) MATLAB Lecture I 17 / 46
Variables

Naming Conventions
Have not to be previously declared
Variable names can contain up to 63 characters
To create a variable, simply assign a value to a name
>> var1 = 1903;
>> myStrings = ’merhaba’;

Variable names
first character must be LETTER
after that, any combination of letters, numbers and
allowable: NetCost, Left2Pay, X3, BJK1903
not allowable: Net-Cost, 1903BJK, %x, @sign

Variable names are case sensitive (var1 is different from Var1)

H. Yücel (METU) MATLAB Lecture I 18 / 46

Variables

Avoid to use built–in variables such as

ans Default variable name for results
eps Smallest incremental number
pi Value of π
inf Infinity
NaN Not a number e.g. 0/0
i and j represent complex numbers

H. Yücel (METU) MATLAB Lecture I 19 / 46

Scalars

A variable can be given a value explicitly (shows up in workspace!)

>> a = 1903

Or as a function of explicit values and existing variables

>> c = 2.4*24-4*a

To suppress output, and the line with a semicolon ;

>> h = 22/7;

H. Yücel (METU) MATLAB Lecture I 20 / 46

Arrays

Like other programming languages, arrays are an important part of MATLAB

Two types of arrays
matrix of numbers (either double or complex)
cell array of objects (more advanced data structure)

Row vector: comma or space separated values between brackets

>> row = [1 4 6 7]
>> row = [1,4,6,7]

Column vector: semicolon separated values between brackets

>> column = [1.4;2;pi]

Size of a vector: length

>> l = length(column)

H. Yücel (METU) MATLAB Lecture I 21 / 46

Special Vector Constructors

linspace()
>> a = linspace(0,10,5)
a =
0 2.5000 5.0000 7.5000 10.0000

Colon operator (:). The basic syntax is

inital:stepsize:final
>> m = 3:8, r = 0:0.25:1, s=1:-1
m =
3 4 5 6 7 8
r =
0 0.2500 0.5000 0.7500 1.0000
s =
Empty matrix: 1-by-0

logspace (to initialize logarithmically spaced values)

H. Yücel (METU) MATLAB Lecture I 22 / 46
Matrices

Make matrices like vectors

>> A = [5 7 9; 1 -3 -7];

Concatenation of vectors
>> r1 = [2 4];
>> r2 = [3 6];
>> M = [r1; r2];

Concatenation of vectors and matrices. Dimensions and Type must coincide!

>> r1 = [2 4];
>> m1 = [3 6; 8 12];
>> M = [r1; m1];

Getting size of the matrix

>> [r,c] = size(M); % size in each dimension
>> r = size(M,1); c = size(M,2);
>> nd = ndims(M); % number of dimensions

H. Yücel (METU) MATLAB Lecture I 23 / 46

Special Matrices

zeros(m,n) m × n matrix of zeros

ones(m,n) m × n matrix of ones
eye(n) n × n identity matrix
rand(m,n)
m × n matrix of uniformly distributed random numbers in range [0,1]
>> M = rand(2,3)
M = 0.8147 0.1270 0.6324
0.9058 0.9134 0.0975

randn(m,n)
m × n matrix of normally distributed random numbers (mean 0, std. dev. 1))
>> M = randn(2,3)
M = -0.4336 3.5784 -1.3499
0.3426 2.7694 3.0349

H. Yücel (METU) MATLAB Lecture I 24 / 46

Replicating and Concatenating Matrices

repmat
>> X = [1 2;3 4];
>> Y = repmat(X,2,3)
Y = 1 2 1 2 1 2
3 4 3 4 3 4
1 2 1 2 1 2
3 4 3 4 3 4
vertcat
>> v1 = [2 3 4]; v2 = [1 2 3];
>> X = vertcat(v1,v2)
X = 2 3 4
1 2 3
horzcat
>> v1 = [2; 3; 4]; v2 = [1; 2; 3];
>> X = horzcat(v1,v2)
X = 2 1
3 2
4 3

H. Yücel (METU) MATLAB Lecture I 25 / 46

Reshaping Matrices

Using the : operator

>> x = round(10*rand(2,4));
>> y = x(:); (The elements of x are stacked in a
column vector, column after column)

reshape()
>> x2 = reshape(y,2,4);
>> M = reshape(linspace(11,18,8),[2,2,2])
M(:,:,1) =
11 13
12 14
M(:,:,2) =
15 17
16 18

H. Yücel (METU) MATLAB Lecture I 26 / 46

Vector Indexing

MATLAB indexing starts with 1, not 0

a(n) returns nth element

The index argument can be vector. In this case, each element is looked up
individually, and returned as a vector of the same size as the index vector.
>> x = [4 6 7 -1 0];
>> a = x(2:4); --------> a=[6 7 -1];
>> b = x(1:end-2); ----> b=[4 6 7];

H. Yücel (METU) MATLAB Lecture I 27 / 46

Matrix Indexing

using subscripts (row and column)

>> A = [1:3;4:6;7:9];
>> A(1:2,:)
ans =
1 2 3
4 5 6
>> A([3 1], [2 3])
ans =
8 9
2 3
>> A([1:2],:) = [] % delete row 1 and 2
A =
7 8 9
using linear indices (as if matrix is vector)
>> [A(2), A(4), A(9)]
ans =
4 2 9

H. Yücel (METU) MATLAB Lecture I 28 / 46

Matrix Indexing

To select rows and columns of a matrix

>> c = [1 4; 0 2];
>> d = c(1,:)
d =
1 4

To get the min. (or max.) value and its index

>> a = [ 1 -1 0 -4, 21];
>> [minVal,minInd] = min(a)
minVal = -4 minInd = 4

To find any indices of specific values or ranges

>> ind = find(a==0);
>> ind = find(a > 0 & a < 4);

To convert between subscripts and indices, use ind2sub and sub2ind

H. Yücel (METU) MATLAB Lecture I 29 / 46

1 Introduction to MATLAB

2 Data Structures: Arrays and Matrices

3 Operators

4 Sparse Matrices

H. Yücel (METU) MATLAB Lecture I 30 / 46

Operations

Arithmetic operations (+,-,*,/)

>> 7/45
>> (2+i)*4/5

Exponentiation (∧ )
>> (3+2*j)^2

Complicated expressions, use parentheses

>> ((2+3)*3)^0.5

Multiplication is NOT implicit given parenthesis

>> 3(1+0.7)
??? 3(1+0.7)
|
Error: Unbalanced or unexpected parenthesis or bracket.

MATLAB has an enormous library of built-in-functions

>> sqrt(2), log(2), log(10)(0.23), cos(pi), atan(2.5)
>> exp(1903), round(1.4), floor(3.3), ceil(4.23)

H. Yücel (METU) MATLAB Lecture I 31 / 46

Transpose

The transpose operators turns a column vector into a row vector and vice
versa
The ′ gives the Hermitian-transpose, i.e., transposes and conjugates all
complex numbers
For vectors of real numbers .′ and ′ give same result
>> a = [ 1;5; 3i+2]
>> a’
ans =
1.0000 5.0000 2.0000 - 3.0000i
>> transpose(a)
ans =
1.0000 5.0000 2.0000 + 3.0000i
>> a.’
ans =
1.0000 5.0000 2.0000 + 3.0000i

H. Yücel (METU) MATLAB Lecture I 32 / 46

Element-Wise Functions

All functions that work on scalars also works on vectors

>> t = [1, pi, 0];
>> f = exp(t);
>> f = [exp(1) exp(pi) exp(0)];
To do element-wise operations, use the dot: .∗, ./, .∧ . Both dimensions must
match (unless one is scalar)
>> u=1:2:8, v=u.^2, w=u./v
u =
1 3 5 7
v =
1 9 25 49
w =
1.0000 0.3333 0.2000 0.1429
>> A = [ 5 7 9; 1 -3 -7]; B = [-1 2 5; 9 0 5];
>> A.*B
ans =
-5 14 45
9 0 -35

H. Yücel (METU) MATLAB Lecture I 33 / 46

Rational and Logical Operators

Boolean values: zero is false, nonzero is true

Some of the logical operators:
Operator Meaning
<, <=, >, >= less than, less than or equal to, etc.
==, ∼= equal to, not equal to
& logical AND
| logical OR
∼ logical NOT
all all true
any any true
xor Xor

H. Yücel (METU) MATLAB Lecture I 34 / 46

Logical Indexing

Construct a matrix R
>> R = rand(5)
R =
0.8147 0.0975 0.1576 0.1419 0.6557
0.9058 0.2785 0.9706 0.4218 0.0357
0.1270 0.5469 0.9572 0.9157 0.8491
0.9134 0.9575 0.4854 0.7922 0.9340
0.6324 0.9649 0.8003 0.9595 0.6787
Test for some logical cases
>> R(R<0.15)’
ans =
0.1270 0.0975 0.1419 0.0357
>> isequal(R(R<0.15), R(find(R<0.15)))
ans =
1

H. Yücel (METU) MATLAB Lecture I 35 / 46

Find

find returns indices of nonzero values. It can simplify code and help avoid
loops
basic syntax: index = find(condition)
>> x = rand(1,10)
x =
Columns 1 through 5
0.4505 0.0838 0.2290 0.9133 0.1524
Columns 6 through 10
0.8258 0.5383 0.9961 0.0782 0.4427
>> inds = find(x>0.4 & x<0.7)
inds =
1 7 10
>> x(inds)
ans =
0.4505 0.5383 0.4427
H. Yücel (METU) MATLAB Lecture I 36 / 46
1 Introduction to MATLAB

2 Data Structures: Arrays and Matrices

3 Operators

4 Sparse Matrices

H. Yücel (METU) MATLAB Lecture I 37 / 46

Dense Matrices

Dense matrix is a matrix in which most of its elements are nonzero.

Any classical approach to create a matrix results a dense matrix in MATLAB.

[,] creates a single row matrix

[;] creates a singe column matrix

zeros(n) returns an n × n matrix of zeros

ones(n) returns an n × n matrix of 1s

diag() creates diagonal matrix of given vector

H. Yücel (METU) MATLAB Lecture I 38 / 46

Create a 1000 × 1000 matrix A
 
−2 1
 
 
 1 −2 1 
 
 .. .. .. 

 . . . 

 
1 −2 1 
 

 
1 −2

M = 1000;
A = diag(ones(M-1,1),-1) + diag(-2*ones(M,1),0) + diag(ones(M-1,1),1);

Compute how much storage this dense matrix need

s = whos(’A’);
by = s.bytes;
>> by = 8000000 bytes
H. Yücel (METU) MATLAB Lecture I 39 / 46
Sparse Matrices

A sparse matrix is a matrix which has relatively small number of nonzero

elements.

Triplet Format in MATLAB stores values and their corresponding row and
column values.

 
row = [1 2 3 1 5 4 1 5];
2 0 2 0 7
col = [1 1 2 3 3 4 5 5];  
8 0 0 0 0
val = [2 8 9 2 4 5 7 3];  
 
S =
0 9 0 0 0
S = sparse(row,col,val);  
0 0 0 5 0
 
 
0 0 4 0 3

H. Yücel (METU) MATLAB Lecture I 40 / 46

spalloc() creates an all zero allocation for a sparse matrix.

m = 10; % number of rows

n = 10; % number of columns
nz = 21; % number of nonzero entries
S = spalloc(m,n,nz);

spones() generates a matrix of 1s with same sparsity structure as matrix S

M = spones(S);

speye() constructs a sparse identity matrix of size m × n

I = speye(m,n);

H. Yücel (METU) MATLAB Lecture I 41 / 46

spdiags() extracts or constructs sparse diagonal matrices.
Extracts nonzero diagonal entries from matrix S

B = spdiags(S);

Extracts diagonals of S specified by d

B = spdiags(S,d);

Replaces the diagonals of S specified by d with columns of B

S = spdiags(B,d,S);

Create m × n sparse matrix from the columns of B and place them along the
diagonals specified by d

S = spdiags(B,d,m,n)

H. Yücel (METU) MATLAB Lecture I 42 / 46

Create a 1000 × 1000 matrix S
 
−2 1
 
 
 1 −2 1 
 
 .. .. .. 

 . . . 

 
1 −2 1 
 

 
1 −2

M = 1000;
S = spdiags([ones(M,1) -2*ones(M,1) ones(M,1)], [-1 0 1] , M, M);

Compute how much storage this dense matrix need

s = whos(’S’);
by = s.bytes;
>> by = 55976 bytes
H. Yücel (METU) MATLAB Lecture I 43 / 46
full() converts a sparse matrix to a dense matrix
A = full(S)
spy() plots sparsity structure of a matrix.
spy(S)

10

20

30

40

50

60

70

80

90

100
0 10 20 30 40 50 60 70 80 90 100
nz = 298

H. Yücel (METU) MATLAB Lecture I 44 / 46

Do not change sparsity structure

Indexing in a sparse structures is a expensive procedure

Accessing the row and column indexes i, j and changing previous value
S(i, j) = c is required

Accessing values is slow in sparse matrices

When an element S(i, j) is requested, a search trough row and column values
is needed

H. Yücel (METU) MATLAB Lecture I 45 / 46

 ...thank you for your attention !

H. Yücel (METU) MATLAB Lecture I 46 / 46