Very Basic MATLAB

Peter J. Olver Marcel Arndt

September 10, 2007

1 Matrices
Type your matrix as follows. Use , or space to separate entries, and ; or
return after each row.
>> A = [4 5 6 -9;5 0 -3 6;7 8 5 0; -1 4 5 1]
or
>> A = [4,5,6,-9;5,0,-3,6;7,8,5,0;-1,4,5,1]
or
>> A = [ 4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1 ]
The output will be:
A =
4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1
You can identify an entry of a matrix by
>> A(2,3)
ans =
-3
A colon : indicates all entries in a row or column
>> A(2,:)
ans =
5 0 -3 6
>> A(:,3)
ans =

1
 6
-3
5
5
You can use these to modify entries
>> A(2,3) = 10
A =
4 5 6 -9
5 0 10 6
7 8 5 0
-1 4 5 1
or to add in rows or columns
>> A(5,:) = [0 1 0 -1]
A =
4 5 6 -9
5 0 10 6
7 8 5 0
-1 4 5 1
0 1 0 -1
or to delete them
>> A(:,2) = []
A =
4 6 -9
5 10 6
7 5 0
-1 5 1
0 0 -1

Accessing Part of a Matrix

>> A = [4,5,6,-9;5,0,-3,6;7,8,5,0;-1,4,5,1]
A =
4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1
>> A([1 3],:)
ans =
4 5 6 -9
7 8 5 0
>> A(:,2:4)
ans =

2
 5 6 -9
0 -3 6
8 5 0
4 5 1
>> A(2:3,1:3)
ans =
5 0 -3
7 8 5

Switching two rows in a matrix

>> A([3 1],:) = A([1 3],:)
A =
7 8 5 0
5 0 -3 6
4 5 6 -9
-1 4 5 1

Special matrices
Zero matrix:
>> zeros(2,3)
ans =
0 0 0
0 0 0
>> zeros(3)
ans =
0 0 0
0 0 0
0 0 0
Identity Matrix:
>> eye(3)
ans =
1 0 0
0 1 0
0 0 1
Matrix of Ones:
>> ones(2,3)
ans =
1 1 1
1 1 1

Random Matrix:

3
>> A = rand(2,3)
A =
0.9501 0.4860 0.4565
0.2311 0.8913 0.0185
Note that the random entries all lie between 0 and 1.

Transpose of a Matrix
>> A = [4,5,6,-9;5,0,-3,6;7,8,5,0;-1,4,5,1]
A =
4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1
>> transpose(A)
ans =
4 5 7 -1
5 0 8 4
6 -3 5 5
-9 6 0 1
>> A
ans =
4 5 7 -1
5 0 8 4
6 -3 5 5
-9 6 0 1

Diagonal of a Matrix
>> diag(A)
ans =
4
0
5
1

Vectors
Vectors are matrices of size 1 along one dimension.
Row vector:
>> v = [1 2 3 4 5]
v =
1 2 3 4 5
Column vector:

4
>> v = [1;2;3;4;5]
v =
1
2
3
4
5
or use transpose operation
>> v = [1 2 3 4 5]
v =
1
2
3
4
5

Forming Other Vectors

>> v = 1:5
v =
1 2 3 4 5
>> v = 10:-2:0
v =
10 8 6 4 2 0
>> v = linspace(0,1,6)
v =
0 0.2000 0.4000 0.6000 0.8000 1.0000

Important: To avoid output, particularly of large matrices, use a semicolon ;

at the end of the line:

>> v = linspace(0,1,100);

gives a row vector whose entries are 100 equally spaced points from 0 to 1.

Size of a Matrix
>> A = [4 5 6 -9 7;5 0 -3 6 -2;7 8 5 0 5 ; -1 4 5 1 -9 ]
A =
4 5 6 -9 7
5 0 -3 6 -2
7 8 5 0 5
-1 4 5 1 -9
>> size(A)
ans =

5
 4 5
>> [m,n] = size(A)
m =
4
n =
5
>> size(A,1)
ans =
4
>> size(A,2)
ans =
5

2 Output Formats
The command format is used to change output format. The default is

>> format short

>> pi
ans =
3.1416
>> format long
>> pi
ans =
3.14159265358979
>> format rat
>> pi
ans =
355/113

This allows you to work in rational arithmetic and gives the best rational
approximation to the answer. Lets return to the default.
>> format short
>> pi
ans =
3.1416

3 Arithmetic operators
+ Matrix addition.
A + B adds matrices A and B. The matrices A and B must have the same
dimensions unless one is a scalar ( 1 1 matrix). A scalar can be added to
anything.

6
>> A = [4,5,6,-9;5,0,-3,6;7,8,5,0;-1,4,5,1]
A =
4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1
>> B = [9 2 4 -9;1 4 -2 -6;8 1 7 0; -3 -4 5 9 ]
B =
9 2 4 -9
1 4 -2 -6
8 1 7 0
-3 -4 5 9
>> A + B
ans =
13 7 10 -18
6 4 -5 0
15 9 12 0
-4 0 10 10

- Matrix subtraction.
A - B subtracts matrix A from B. Note that A and B must have the same
dimensions unless one is a scalar.
>> A - B
ans =
-5 3 2 0
4 -4 -1 12
-1 7 -2 0
2 8 0 -8

Scalar multiplication
>> 3*A - 4*B
ans =
-24 7 2 9
11 -16 -1 42
-11 20 -13 0
9 28 -5 -33

Matrix multiplication.
A*B is the matrix product of A and B. A scalar (a 1-by-1 matrix) may multiply
anything. Otherwise, the number of columns of A must equal the number of
rows of B.
>> A * B

7
ans =
116 70 3 -147
3 -17 29 9
111 51 47 -111
32 15 28 -6
Note that two matrices must be compatible before we can multiply them.
The order of multiplication is important!
>> v = [1 2 3 4]
v =
1 2 3 4
>> w = [1;2;3;4]
w =
1
2
3
4
>> v * w
ans =
30
>> w * v
ans =
1 2 3 4
2 4 6 8
3 6 9 12
4 8 12 16

. Array multiplication
A.*B denotes element-by-element multiplication. A and B must have the same
dimensions unless one is a scalar. A scalar can be multiplied into anything.
>> a = [3 4 5 6 7 8 9]
a =
3 4 5 6 7 8 9
>> b = [8 6 2 4 5 6 -1]
b =
8 6 2 4 5 6 -1
>> a .* b
ans =
24 24 10 24 35 48 -9

Matrix power.
C = An is A to the n-th power if n is a scalar and A is square. If n is an integer
greater than one, the power is computed by repeated multiplication.

8
>> A = [4 5 6 -9;5 0 -3 6;7 8 5 0; -1 4 5 1 ]
A =
4 5 6 -9
5 0 -3 6
7 8 5 0
-1 4 5 1
>> A^3
ans =
501 352 351 -651
451 169 -87 174
1103 799 533 -492
445 482 413 -182

. Array power.
C = A.B denotes element-by-element powers. A and B must have the same
dimensions unless one is a scalar. A scalar can go in either position.
>> A = [8 6 2 4 5 6 -1 ]
A =
8 6 2 4 5 6 -1
>> A.^3
ans =
512 216 8 64 125 216 -1

Length of a Vector, Norm of a Vector, Dot Product

>> u = [8 -7 6 5 4 -3 2 1 9]
u =
8 -7 6 5 4 -3 2 1 9
>> length(u)
ans =
9
>> norm(u)
ans =
16.8819
>> v = [9 -8 7 6 -4 5 0 2 -4]
v =
9 -8 7 6 -4 5 0 2 -4
>> dot(u,v)
ans =
135
>> u*v
ans =
135

9
4 Complex Numbers
>> u = [2-3i, 4+6i,-3,+2i]
u =
2.0000- 3.0000i 4.0000+ 6.0000i -3.0000 0+ 2.0000i
>> conj(u)
ans =
2.0000+ 3.0000i 4.0000- 6.0000i -3.0000 0- 2.0000i

Hermitian transpose:
>> u
ans =
2.0000+ 3.0000i
4.0000- 6.0000i
-3.0000
0- 2.0000i
Other operations:
>> norm(u)
ans =
8.8318
>> dot(u,u)
ans =
78
>> sqrt(ans)
ans =
8.8318
>> u*u
ans =
78

5 Solving Systems of Linear Equations

The best way of solving a system of linear equations

Ax = b

in MatLab is to use the backslash operation \ (backwards division)

>> A = [1 2 3;-1 0 2;1 3 1]

A =
1 2 3
-1 0 2
1 3 1
>> b = [1; 0; 0]

10
b =
1
0
0
>> x = A \ b
x =
0.6667
-0.3333
0.3333

The backslash is implemented by using Gaussian elimination with partial piv-

oting. An alternative, but less accurate, method is to compute inverses:

>> B = inv(A)
B =
0.6667 -0.7778 -0.4444
-0.3333 0.2222 0.5556
0.3333 0.1111 -0.2222

or
>> B = A^(-1)
B =
0.6667 -0.7778 -0.4444
-0.3333 0.2222 0.5556
0.3333 0.1111 -0.2222
>> x = B * b
x =
0.6667
-0.3333
0.3333
Another method is to use the command rref:
To solve the following system of linear equations:

x1 + 4x2 2x3 + x4 = 2
2x1 + 9x2 3x3 2x4 = 5
x1 + 5x2 x4 = 3
3x1 + 14x2 + 7x3 2x4 = 6

we form the augmented matrix:

>> A = [1,4,-2,3,2; 2,9,-3,-2,5; 1,5,0,-1,3; 3,14,7,-2,6]
A =
1 4 -2 3 2
2 9 -3 -2 5
1 5 0 -1 3

11
 3 14 7 -2 6
>> rref(A)
ans =
1.0000 0 0 0 -5.0256
0 1.0000 0 0 1.6154
0 0 1.0000 0 -0.2051
0 0 0 1.0000 0.0513
The solution is: x1 = 5.0256, x2 = 1.6154, x3 = 0.2051, x4 = 0.0513.
Case 1: Infinitely many solutions:

>> A = [-2 2 -2;1 -1 1; 2 -2 2]

A =
-2 2 -2
1 -1 1
2 -2 2
>> b = [-8; 4; 8]
b =
-8
4
8
>> A \ b
Warning: Matrix is singular to working precision.
ans =
NaN
NaN
NaN

MatLab is unable to find the solutions. In this case, we can apply rref to the
augmented matrix.
>> C = [A b]
C =
-2 2 -2 -8
1 -1 1 4
2 -2 2 8
>> rref(C)
ans =
1 -1 1 4
0 0 0 0
0 0 0 0
Conclusion: There are infinitely many solutions since row 2 and row 3 are all
zeros.
Case 2: No solutions:
>> A = [-2 1; 4 -2]
A =

12
 -2 1
4 -2
>> b = [5; -1]
b =
5
-1
>> A \ b
Warning: Matrix is singular to working precision.
ans =
Inf
Inf
>> C = [A b]
C =
-2 1 5 4 -2 -1
>> rref(C)
ans =
1.0000 -0.5000 0
0 0 1.0000
Conclusion: Row 2 is not all zeros, and the system is incompatible.
Important: If the coefficient matrix A is rectangular (not square) then A \ b
gives the least squares solution (relative to the Euclidean norm) to the system
A x = b. If the solution is not unique, it gives the least squares solution x with
minimal Euclidean norm.

>> A = [1 1;2 1;-5, -1]

A =
1 1
2 1
-5 -1
>> b = [1;1;1]
b =
1
1
1
>> A \ b
ans =
-0.5385
1.7692

If you want the least squares solution in the square case, one trick is to add an
extra equation 0 = 0 to make the coefficient matrix rectangular:

>> A = [-2 2 -2;1 -1 1; 2 -2 2]

A =
-2 2 -2
1 -1 1

13
 2 -2 2
>> b=[-8; 4; 8]
b =
-8
4
8
>> A \ b
Warning: Matrix is singular to working precision.
ans =
Inf
Inf
Inf
>> A(4,:) = 0
A =
-2 2 -2
1 -1 1
2 -2 2
0 0 0
>> b(4) = 0
b =
-8
4
8
0
>> A \ b
Warning: Rank deficient, rank = 1 tol = 2.6645e-15.
ans =
4.0000
0
0

6 Plotting Functions
Functions can be stored as vectors. Namely, a vector x and a vector y of the
same length correspond to the sampled function values (xi , yi ).
To plot the function y = x2 .5 x first enter an array of independent variables:
>> x = linspace(0,1,25)
>> y = x.^2 - .5*x;
>> plot(x,y)
The plot shows up in a new window. To plot in a different color, use
>> plot(x,y,r)
where the character string r means red. Use the helpwindow to see other
options.

14
To plot graphs on top of each other, use hold on.
>> hold on
>> z = exp(x);
>> plot(x,z)
>> plot(x,z,g)
hold off will stop simultaneous plotting. Alternatively, use
>> plot(x,y,r,x,z,g)

Surface Plots
Here x and y must give a regtangular array, and z is a matrix whose entries are
the values of the function at the array points.

>> x =linspace(-1,1,40); y = x;
>> z = x * (y.^2);
>> surf(x,y,z)

Typing the command

>> rotate3d

will allow you to use the mouse interactively to rotate the graph to view it from
other angles.

7 Functions, Subroutines, M-Files

Simple functions can be declared as anonymous functions:
>> f = @(x) 1./x
f =
@(x)1./x
>> f(5)
ans =
0.2000
>> f(1:5)
ans =
1.0000 0.5000 0.3333 0.2500 0.2000
For more complex functions or subroutines, use M-Files. Create a file with the
name of the subroutine and the suffix .m. For the trapezoidal rule
" n1
#
h X
Tn := f (a) + 2 f (a + ih) + f (b)
2 i=1

use

15
>> edit trapezoidal.m
An editor pops up. Enter the code of the subroutine:
function integral = trapezoidal(f, a, b, n)
h = (b-a)/n;
integral = 0;
for i=1:(n-1)
integral = integral + f(a+i*h);
end
integral = 0.5 * h * ( f(a) + 2*integral + f(b) );
Save the file.
>> trapezoidal(f, 1, 2, 4)
ans =
0.6970
The subroutine gets much faster for large numbers n by avoiding the loop in
your M-File. Save this code as trapezoidal2.m:
function integral = trapezoidal2(f, a, b, n)
h = (b-a)/n;
integral = sum(f(a+(1:(n-1))*h));
integral = 0.5 * h * ( f(a) + 2*integral + f(b) );
Multiple values can be returned as follows. Each entry can be a matrix itself.
Save this code as multreturn.m:

function [a,b,c] = multreturn(x,y)

a = x+y;
b = x-y;
c = rand(2,3);
Then

>> [u,v,w] = multreturn(6,2)

u =
8
v =
4
w =
0.6557 0.8491 0.6787
0.0357 0.9340 0.7577

16