First Steps with Scilab Vector and Matrix

Lecture 1: Introduction to Scilab

Ahmed Kebaier
[email protected]

HEC, Paris
First Steps with Scilab Vector and Matrix

Outline

1 First Steps with Scilab

2 Vector and Matrix

First Steps with Scilab Vector and Matrix

Outline

1 First Steps with Scilab

2 Vector and Matrix

First Steps with Scilab Vector and Matrix

After launching Scilab, you can test the following commands

−− > help // To open Scilab’s help
−− > help mot-clé // Get a description of the Keyword
−− > apropos mot-clé // Get pages related to the Keyword
−− > quit // To quit Scilab
First Steps with Scilab Vector and Matrix

After launching Scilab, you can test the following commands

−− > help // To open Scilab’s help
−− > help mot-clé // Get a description of the Keyword
−− > apropos mot-clé // Get pages related to the Keyword
−− > quit // To quit Scilab

The command-line interface clear, clc et clf erase

respectively data in memory, the commands on the screen and the
plots. These commands should be executed regularly to avoid
errors and make the memory free.
First Steps with Scilab Vector and Matrix

After launching Scilab, you can test the following commands

−− > help // To open Scilab’s help
−− > help mot-clé // Get a description of the Keyword
−− > apropos mot-clé // Get pages related to the Keyword
−− > quit // To quit Scilab

The command-line interface clear, clc et clf erase

respectively data in memory, the commands on the screen and the
plots. These commands should be executed regularly to avoid
errors and make the memory free.

Test the following command-line on the interface:

−− > x=1
−− > A=ones(3,4)
−− > x+A
First Steps with Scilab Vector and Matrix

Often a file entitled for example test.sce, containing the

following instructions :
clc;clf;clear;
A=ones(3,4)
1+A
First Steps with Scilab Vector and Matrix

Often a file entitled for example test.sce, containing the

following instructions :
clc;clf;clear;
A=ones(3,4)
1+A
Under Scilab, tape : −− > exec(“test.sce”)
We can also create a function “*.sci”. For this,
1 Open a file entitled for example carre.sci, containing the

following instructions :
function d = carre(x)
d= x.*x
endfunction
2 Under Scilab load and execute the file carre.sci :

−− > getf(“carre.sci”) // if the file is in the current

folder.
Function carre is now defined under scilab :
−− > x=[0,1,2,3,4]
−− > carre(x)
First Steps with Scilab Vector and Matrix

Exercise

Plot the cos function on the interval [−π, π]. Create a mesh of 11
points.
First Steps with Scilab Vector and Matrix

Exercise

Plot the cos function on the interval [−π, π]. Create a mesh of 11
points.

Solution: Enter the following lines in the command window

x=linspace(-%pi,%pi,11);
y=cos(x)./(1+x.^ 2);
clf()
plot(x,y,b)
First Steps with Scilab Vector and Matrix

Outline

1 First Steps with Scilab

2 Vector and Matrix

First Steps with Scilab Vector and Matrix

The easiest way to define a matrix n × m under Scilab is to write

all its components using the keyboard:

A = [a1,1 , . . . , a1,m ; . . . ; an,1 , . . . , an,m ]

First Steps with Scilab Vector and Matrix

The easiest way to define a matrix n × m under Scilab is to write

all its components using the keyboard:

A = [a1,1 , . . . , a1,m ; . . . ; an,1 , . . . , an,m ]

Elementary operations. Test on examples !

−− > A+B //sum

−− > A*B //product
−− > A.*B //operations are performed component wise
−− > Aˆ 2 //equivalent to A*A
−− > A.ˆ 2 //equivalent to A.*A
−− > det(A) //determinent of A
−− > A’ //transpose of A
−− > inv(A) //inverse of A
First Steps with Scilab Vector and Matrix

if A is a matrix but not a vector diag(A,k) extracts the

diagonal number k as a column vector.
A*B is the matrix product of the matrices A and B (or product
between a scalar and a vector or matrix).
A’ performs the transposition of matrix A.
If A is a square invertible matrix you can solve the linear
system Ax = b using x = A\ b
(a P A = LU factorization of the matrix, followed by an estimation
of its condition number, and finally by solving the 2 triangular
systems, are done in a transparent manner)
First Steps with Scilab Vector and Matrix

Exercise
give a value to the variable n then define the n × n matrix :
 
2 −1 0 . . . 0
−1 . . . . . . . . .
 .. 
 . 

A =  0 ... ... ... 0 
 
 ..
 
.. .. ..
. . . −1

 .
0 . . . 0 −1 2

Define a vector b ∈ Rn using rand(n,1), compute the solution

of Ax = b. Compute the relative residual kAx − bk/kbk using the
function norm.
Compute E = 21 x T Ax − b T x. Define another vector y ∈ Rn
using rand, compute F = 12 y T Ay − b T y and verify that E < F .
 
A In
Build the following matrix: B = A =
In A
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
// another solution
// A = diag(v,-1) + diag(2*ones(1,n)) + diag(v,1)
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
// another solution
// A = diag(v,-1) + diag(2*ones(1,n)) + diag(v,1)
b = rand(n,1);
x = A\ b
res = norm(A*x-b)/norm(b)
E=0.5*x’*A*x - b’*x
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
// another solution
// A = diag(v,-1) + diag(2*ones(1,n)) + diag(v,1)
b = rand(n,1);
x = A\ b
res = norm(A*x-b)/norm(b)
E=0.5*x’*A*x - b’*x
y=rand(n,1);
F=0.5*y’*A*y - b’*y
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
// another solution
// A = diag(v,-1) + diag(2*ones(1,n)) + diag(v,1)
b = rand(n,1);
x = A\ b
res = norm(A*x-b)/norm(b)
E=0.5*x’*A*x - b’*x
y=rand(n,1);
F=0.5*y’*A*y - b’*y
E<F
First Steps with Scilab Vector and Matrix

Solution

n = 7;
v = -ones(1,n-1);
A = diag(v,-1) + 2*eye(n,n) + diag(v,1)
// another solution
// A = diag(v,-1) + diag(2*ones(1,n)) + diag(v,1)
b = rand(n,1);
x = A\ b
res = norm(A*x-b)/norm(b)
E=0.5*x’*A*x - b’*x
y=rand(n,1);
F=0.5*y’*A*y - b’*y
E<F
B = [ A , eye(n,n) ;... eye(n,n), A ]
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
A(2,:) // change row 2 of A
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
A(2,:) // change row 2 of A
A(2,:) = ones(1,4)
// change column 3 of A
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
A(2,:) // change row 2 of A
A(2,:) = ones(1,4)
// change column 3 of A
A(:,3) = 0 // extract submat (1,3)x(1,2) and assign
it to B
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
A(2,:) // change row 2 of A
A(2,:) = ones(1,4)
// change column 3 of A
A(:,3) = 0 // extract submat (1,3)x(1,2) and assign
it to B
B = A([1,3],[1 2])
// change the same sub-matrix
First Steps with Scilab Vector and Matrix

Assignments and extractions

Try: A = rand(3,4) // create a matrix

// change coef (2,2) of A
A(2,2) = -1
// extract coef (2,3) of A and assign it to variable
c
c = A(2,3)
// extract row 2 of A (it is assigned to ans)
A(2,:) // change row 2 of A
A(2,:) = ones(1,4)
// change column 3 of A
A(:,3) = 0 // extract submat (1,3)x(1,2) and assign
it to B
B = A([1,3],[1 2])
// change the same sub-matrix
A([1,3],[1 2]) = [-10,-20;-30,-40]
First Steps with Scilab Vector and Matrix

Another very useful vector constructor In the

command window

Try the following expressions:

I = 1:5
J = 1:2:6
// try also J = 1:2:7 which give the same vector
K = 10:-1:5
The syntax is init val:inc:lim and this builds a row vector
with init val as the first coefficient, the others components being
obtained from the previous one by adding it inc until lim is not
overtaken.
First Steps with Scilab Vector and Matrix

Exercise 3

1 Copy-paste your previous script exercise2.sce in a new file

exercise3.sce and use a small value for n (e.g. n = 5) (Rmk:
we need only the part of the code which defines A and B: you
can remove unuseful lines of code).
2 Continue the script by creating the following new matrices:
1 C such that Ci,j = Ai,n+1−j i.e. by reversing the column order
of A;
2 D such that Ci,j = A2i−1,j i.e. taking one row over two of
matrix A;
3 E the matrix formed by the B submatrix of rows and columns
n − 2, n − 1, n, n + 1, n + 2, n + 3.
First Steps with Scilab Vector and Matrix

Solution

n=5;
v=-ones(1,n-1);
A=diag(v,-1) + 2*eye(n,n) + diag(v,1)
B=[ A , eye(n,n) ;... eye(n,n), A ]
First Steps with Scilab Vector and Matrix

Solution

n=5;
v=-ones(1,n-1);
A=diag(v,-1) + 2*eye(n,n) + diag(v,1)
B=[ A , eye(n,n) ;... eye(n,n), A ]
C = A(:,n:-1:1)
First Steps with Scilab Vector and Matrix

Solution

n=5;
v=-ones(1,n-1);
A=diag(v,-1) + 2*eye(n,n) + diag(v,1)
B=[ A , eye(n,n) ;... eye(n,n), A ]
C = A(:,n:-1:1)
D = A(1:2:n,:)
First Steps with Scilab Vector and Matrix

Solution

n=5;
v=-ones(1,n-1);
A=diag(v,-1) + 2*eye(n,n) + diag(v,1)
B=[ A , eye(n,n) ;... eye(n,n), A ]
C = A(:,n:-1:1)
D = A(1:2:n,:)
E = B(n-2:n+3,n-2:n+3)
First Steps with Scilab Vector and Matrix

The component-wise algebra

Three useful operators .*, ./ and .ˆ :

1 x and y matrices with the same dimensions:

z=x.*y is the component-wise product, i.e. zi,j = xi,j yi,j

z=x./y is the component-wise product, i.e. zi,j = xi,j /yi,j
Useful shortcut: if s is a scalar, z=s./y gives zi,j = si,j /yi,j
but z = 1./y doesn’t work as expected ! (use z = 1 ./y ).
2 x matrix and p scalar:
p
z=x.^ p is the component-wise power: zi,j = xi,j .
z=p.^ x is the component-wise power: zi,j = p xi,j .
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
clf(); // clear the graphic window
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
clf(); // clear the graphic window
subplot(2,1,1); // split the graphic window and use
subpart 1
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
clf(); // clear the graphic window
subplot(2,1,1); // split the graphic window and use
subpart 1
plot(x,y1,"ro",x,y2,"bx"); // only symbols
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
clf(); // clear the graphic window
subplot(2,1,1); // split the graphic window and use
subpart 1
plot(x,y1,"ro",x,y2,"bx"); // only symbols
subplot(2,1,2); // split the graphic window and use
subpart 2
First Steps with Scilab Vector and Matrix

The plot function I

x = linspace(0,2*%pi,31);
y1 = sin(x); y2 = cos(x);
scf(0); // select graphic window 0 to be the default
graphic window
clf(); // clear the graphic window
plot(x,y1,"b-",x,y2,"r--"); // only lines
scf(1); // select graphic window 1 to be the default
graphic window
clf(); // clear the graphic window
subplot(2,1,1); // split the graphic window and use
subpart 1
plot(x,y1,"ro",x,y2,"bx"); // only symbols
subplot(2,1,2); // split the graphic window and use
subpart 2
plot(x,y1,"r--o",x,y2,"g-x"); // both lines and
symbols
First Steps with Scilab Vector and Matrix

The plot function II

1 A title with title(string title).

2 x and y labels with xlabel(string xlabel) and
ylabel(string ylabel)
3 A legend for the curves with legend(curve1 leg, curve2
leg)
First Steps with Scilab Vector and Matrix

Programming tools I

Functions. In scilab a function definition takes the form:

function [y1,y2,...,yn] = function name(x1,x2,..xm)

// the body of the function define the output
arguments y1,...,yn
// in function of the input arguments x1,...,xm
endfunction

Such a definition can be written in a script (before the part of the

script which uses it) or better in another file (with a name
traditionaly ending with .sci). You can write any number of
functions in a file. In this case you have to load the file in scilab
before we can use them.
First Steps with Scilab Vector and Matrix

Programming tools II
if tests
They permit to execute different blocks of code depending on
boolean expressions:
if bool expression then
// block executed when bool expression is TRUE
.....
else
// block executed when bool expression is FALSE
.....
end
Example
x = rand()
if x < 0.5 then
y = -1;
else
y = 1;
First Steps with Scilab Vector and Matrix

Programming tools III

For loop
for i = row vector
// body of the loop
.....
end
the number of iterations equal the number of components of
the row vector
at iteration k the loop variable i is equal to row vector(k).
Very often the row vector is of the form first val:inc:lim.
First Steps with Scilab Vector and Matrix

Programming tools III

For loop
for i = row vector
// body of the loop
.....
end
the number of iterations equal the number of components of
the row vector
at iteration k the loop variable i is equal to row vector(k).
Very often the row vector is of the form first val:inc:lim.
It is possible to exit prematurely a for loop using the break
statement:
for i = 1:n
.....
if special condition test then, break, end
....
end
First Steps with Scilab Vector and Matrix

Programming tools IV

while loop
A while loop allows to repeat a block of code while a boolean
expression is true:

while bool expression

// block
....
end
First Steps with Scilab Vector and Matrix

Programming tools IV

while loop
A while loop allows to repeat a block of code while a boolean
expression is true:

while bool expression

// block
....
end

Try:

x = 1;
while x < 1000, x = 2*x, end
First Steps with Scilab Vector and Matrix

Programming tools IV

while loop
A while loop allows to repeat a block of code while a boolean
expression is true:

while bool expression

// block
....
end

Try:

x = 1;
while x < 1000, x = 2*x, end

It is also possible to exit prematurely a while loop with the break

statement.