A Short Introduction to Scilab
Terence Leung Ho Yin, Tsing Nam KiuTable of Contents
- About Scilab
- Installing and Running Scilab
- Documentation and Help
- Scilab Basics
- Common Operators
- Common Functions
- Special Constants
- The Command Line
- Data Structures
- Strings
- Saving and Loading Variables
- Dealing with Matrices
- Entering Matrices
- Calculating Sums
- Subscripts
- The Colon Operator
- Simple Matrix Generation
- Concatenation
- Deleting Rows and Columns
- Matrix Inverse and Solving Linear Systems
- Entry-wise operations, Matrix Size
- The Programming Environment
- Creating Functions
- Flow Control
- Some Programming Tips
- Debugging
- Plotting Graphs
- 2D Graphs
- 3D Surfaces
- Scilab versus Matlab
- References
1. About Scilab
Scilab is a freely distributed open source scientific software package, first developed by researchers from INRIA and ENPC, and now by the Scilab Consortium. It is similar to Matlab, which is a commercial product. Yet it is almost as powerful as Matlab. Scilab consists of three main components:
- an interpreter
- libraries of functions (Scilab procedures)
- libraries of Fortran and C routines
Scilab is specialized in handling matrices (basic matrix manipulation, concatenation, transpose, inverse, etc.) and numerical computations. Also it has an open programming environment that allows users to create their own functions and libraries.
For further information and documentation, visit the Scilab homepage:
2. Installing and Running Scilab
First, you must have the software. Go to the download section in the Scilab homepage, find a right version for your operating system (platform), and then click to download. For easy installation, it is advisable to download the installer (for binary version). Then double click the downloaded file and follow the instructions to complete the installation.
To run Scilab, type
scilex
in the command prompt in the folder bin under the installation directory, or click the shortcut in the start menu if you use Windows. To exit the program, type
exit
or close the window of the main program.
3. Documentation and Help
To find the usage of any function, type
help function_name
For example:
help sum
If you want to find functions that you do not know, you may just type
help
and search for the keywords of the functions. Finally, if you want more information, you may visit the Scilab homepage. There is a section called documentation. It is very resourceful.
4. Scilab Basics
Common Operators
Here is a list of common operators in Scilab:
| + | addition |
| - | subtraction |
| * | multiplication |
| / | division |
| ^ | power |
| ' | conjugate transpose |
Common Functions
Some common functions in Scilab are:
sin, cos, tan, asin, acos, atan, abs, min, max, sqrt, sum
E.g., when we enter:
sin(0.5)
then it displays:
ans = 0.4794255
Another example:
max(2, 3, abs(-5), sin(1)) ans = 5
Special Constants
We may wish to enter some special constants like, i (sqrt(-1)) and e. It is done by entering
%pi %i %e
respectively. There are also constants
%t %f
which are Boolean constants representing true and false, respectively. Boolean variables would be introduced later.
The Command Line
Multiple commands, separated by commas, can be put on the same command line:
A = [1 2 3], s = tan(%pi/4) + %e
A =
1. 2. 3.
s =
3.7182818
Entering a semi-colon at the end of a command line suppresses showing the result (the answer of the expression):
A = [1 2 3]; s = tan(%pi/4) + %e;
Here the vector [1 2 3] is stored in the variable A, and the expression tan(%pi/4) + %e is evaluated and stored in s, but the results are not shown on the screen.
A long command instruction can be broken with line-wraps by using the ellipsis (...) at the end of each line to indicate that the command actually continues on the next line:
s = 1 -1/2 + 1/3 -1/4 + 1/5 - 1/6 + 1/7 ...
- 1/8 + 1/9 - 1/10 + 1/11 - 1/12;
Anything that is typed after a pair of slashes // will be ignored, and hence can serve as comments or code annotations:
%e^(%pi * %i) + 1 // should equal 0, as in the Euler's identity ans = 1.225D-16i
Using the up and down arrow in the command line can recall previous commands.
Data Structures
Scilab supports many data structures. Examples are: (real or complex) matrix, polynomial, Boolean, string, function, list, tlist, sparse, library. Please read the Scilab documentation for details. To query for the type of an object, type:
typeof(object)
Strings
To enter a string, enclose it with either single or double quotations. For example:
'This is a string'
or
"this is also a string"
To concatenate strings, use the operator + :
"Welcome " + "to " + "Scilab!" ans = Welcome to Scilab!
There are some basic string handling functions such as
strindex strsplit strsubst part
Please refer to Scilab's documentation for details.
Saving and Loading Variables
To save and to load variables, we use the save and load functions:
save('file_name', var1, var2, ...);
load('file_name', 'var1', 'var2', ...);
where file_name is the name of the file to be saved or loaded, and var1, var2, ..., are variable names.
Notice that the variable name has to match when it is to be saved. Here are some illustrations.
a = 3; b = %f; s = 'scilab';
save('save.dat', a, b, s);
clear a; // delete the variable a
clear b;
clear s;
load('save.dat', 'a', 'b', 's');
// load all the saved variables
load('save.dat','b');
// It loads only variable b, but not
// variable a in the name of b
load('save.dat','d');
// It will not show any error messages.
// Variable d is undefined, not empty.
listvarinfile('save.dat');
// list variables in a file saved by
// the function save
Name Type Size Bytes
----------------------------------------------------
a constant 1 by 1 24
b boolean 1 by 1 20
s string 1 by 1 44
5. Dealing with Matrices
Entering Matrices
There are many ways to enter a matrix. Here is the simplest method:
- Separate each element in a row by a blank space or a comma,
- Separate each row of elements with a semi-colon, and
- Put the whole list of elements in a pair of square brackets.
For example, to enter a 3 x 3 magic square and assign to the variable M :
M = [8 1 6; 3 5 7; 4 9 2] M = 8. 1. 6. 3. 5. 7. 4. 9. 2.
Calculating Sums
For a magic square, we may wish to check for its column sums and row sums and the sum of diagonals. This is done by entering:
sum(M,'c') // column sums ans = 15. 15. 15.
sum(M,'r') // row sums ans = 15. 15. 15.
The sum of the main diagonal is easily done with the help of the function diag.
diag(M) ans = 8. 5. 2.
Subscripts
It is a bit more difficult to find the sum of the other diagonal. We will show two ways to accomplish it. One method is to find the sum manually, i.e., to read the specified elements and then to sum them up.
M(1,3) + M(2,2) + M(3,1) ans = 15.
It is possible to access elements in a matrix using a single index. This by treating a matrix as a long vector formed by stacking up the columns of the matrix. E.g., the values of M(1), M(2), M(3), M(4), M(5) are 8, 3, 4, 1, 5, respectively.
Accessing out-of-bound elements will result in an error, like entering:
M(3,4)
!--error 21
invalid index
A smarter way to get the sum of the other diagonal is to use the function mtlb_fliplr, where mtlb stands for Matlab. This is to flip a matrix left-to-right (lr):
mtlb_fliplr(M) ans = 6. 1. 8. 7. 5. 3. 2. 9. 4.
The desired result would then be obtained by typing sum(diag(mtlb_fliplr(M))).
The Colon Operator
The colon operator is one of the most important operators in Scilab. The expression 1:10 results in a row matrix with elements 1, 2, ..., 10, i.e.:
1:10 ans = 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
To have non-unit spacing we specify the increment:
10 : -2 : 2 ans = 10. 8. 6. 4. 2.
Notice that expressions like 10:-2:1, 10:-2:0.3 would produce the same result while 11:-2:2 would not.
Subscript expressions involving colons refer to parts of a matrix. M(i:j, k) shows the i-th row to j-th row of column k. Similarly,
M(3,2:3) ans = 9. 2.
Some more examples:
M(3,[3,2]) ans = 2. 9. M([2,1], 3:-1:1) ans = 7. 5. 3. 6. 1. 8.
The operator $, which gives the largest value of an index, is handy for getting the last entry of a vector or a matrix. For example, to access all elements except the last of the last column, we type:
M(1:$-1, $)
We sometimes want a whole row or a column. For example, to obtain all the elements of the second row of M, enter:
M(2,:) ans = 3. 5. 7.
Now we have a new way to perform operations like mtlb_fliplr(M). It is done by entering M(:, $:-1:1). However the function mtlb_fliplr(M) would obtain result faster (in computation time) than using the subscript expression.
Simple Matrix Generation
Some basic matrices can be generated with a single command:
| zeros | all zeros |
| ones | all ones |
| eye | dentity matrix (having 1 in the main diagonal and 0 elsewhere) |
| rand | random elements (follows either normal or uniform distribution) |
Some illustrations:
zeros(2,3) ans = 0. 0. 0. 0. 0. 0. 8 * ones(2,2) ans = 8. 8. 8. 8. eye(2,3) ans = 1. 0. 0. 0. 1. 0. rand(1,3,'uniform') // same as rand(1,3) ans = 0.2113249 0.7560439 0.0002211
Concatenation
Concatenation is the process of joining smaller size matrices to form bigger ones. This is done by putting matrices as elements in the bigger matrix:
a = [1 2 3]; b = [4 5 6]; c = [7 8 9]; d = [a b c] d = 1. 2. 3. 4. 5. 6. 7. 8. 9. e = [a; b; c] e = 1. 2. 3. 4. 5. 6. 7. 8. 9.
Concatenation must be row/column consistent:
x = [1 2]; y = [1 2 3];
z = [x; y]
!-error 6
inconsistent row/column dimensions
We can also concatenate block matrices, e.g.:
[eye(2,2) 5*ones(2,3); zeros(1,3) rand(1,2)] ans = 1. 0. 5. 5. 5. 0. 1. 5. 5. 5. 0. 0. 0. 0.6525135 0.3076091
Remember that it is an error to access out-of-bound element of a matrix. However, it is okay to assign values to out-of-bound elements. The result is a larger size matrix with all unspecified entries 0:
M = matrix(1:6, 2, 3); M(3,1) = 10 M = 1. 3. 5. 2. 4. 6. 10. 0. 0.
It is remarked that this method is slow. If the size of the matrix is known beforehand, we should use pre-allocation:
M = zeros(3,3); // pre-allocation M([1 2], :) = matrix(1:6, 2, 3); M(3,1) = 10;
Deleting Rows and Columns
A pair of square brackets with nothing in between represents the empty matrix. This can be used to delete rows or columns of a matrix. To delete the 1st and the 3rd rows of a 4x4 identity matrix, we type:
A = eye(4,4); A([1 3],:) = [] A = 0. 1. 0. 0. 0. 0. 0. 1.
If we delete a single element from a matrix, it results in an error, e.g.:
A(1,2) = []
!-error 15
submatrix incorrectly defined
If we delete elements using single index expression, the result would be a column vector:
B=[1 2 3; 4 5 6]; B(1:2:5)=[] B = 4. 5. 6.
Matrix Inverse and Solving Linear Systems
The command inv(M) gives the inverse of the matrix M. If the matrix is badly scaled or nearly singular, a warning message will be displayed:
inv([1 2;2 4.0000001])
warning
matrix is close to singular or badly scaled. rcond = 2.7778D-09
ans =
40000001. - 20000000.
- 20000000. 10000000.
Solving a system of linear equations Ax = b, i.e., to find x that satisfies the equation, when A is a square, invertible matrix and b is a vector, is done in Scilab by entering A\b :
A = rand(3,3), b = rand(3,1)
A =
0.2113249 0.3303271 0.8497452
0.7560439 0.6653811 0.6857310
0.0002211 0.6283918 0.8782165
b =
0.0683740
0.5608486
0.6623569
x = A \ b
x =
- 0.3561912
1.7908789
- 0.5271342
Another method is to type inv(A) * b. Although it gives the same result, it is slower than A\b because the first method mainly uses Gaussian Elimination which saves some computation effort. Please read the Scilab help file for more details about the slash operator when A is non-square.
A / b solves for x in the equation xb = A.
Entry-wise operations, Matrix Size
To add 4 to each entry of a matrix M, using M + 4 * ones(M) is correct but troublesome. Indeed this can be done easily by M + 4. Subtraction of a scalar from a matrix entry-wise is done similarly.
Multiplying 2 to the second column and 3 to the third column of M can be achieved by using the entry-wise multiplication operator .* :
M .* [1:3; 1:3]
Entry-wise arithmetic operations for arrays are:
| + | addition |
| - | subtraction |
| .* | multiplication |
| .^ | power |
| ./ | right division |
| .\ | left division |
Thus, instead of typing
s = 1 -1/2 + 1/3 -1/4 + 1/5 - 1/6 + 1/7 ...
- 1/8 + 1/9 - 1/10 + 1/11 - 1/12;
one may simply type
s = sum((1:2:12) .\ 1) - sum((2:2:12) .\ 1)
or
s = sum(1 ./ (1:2:12)) - sum(1 ./ (2:2:12))
Note that 1./1:2:12 is interpreted as (1./1):2:12. Similarly, 1:2:12.\1 is interpreted as 1:2:(12.)\1.
To enter the matrix M = [1 2 3 4 5; 6 7 8 9 10], one may use:
M = zeros(2,5); M(:) = 1:10;
Yet an almost effortless method is to use the function matrix, which reshapes a matrix to a desired size:
M = matrix(1:10,5,2)'
How to enter N = [1 2; 3 4; 5 6; 7 8; 9 10] easily? Hint: think of some simple operations on a matrix.
A handy function in Scilab called size returns the dimensions of the matrix in query:
size(M) ans = 2. 5.
while size(M,1) returns 2 (number of rows) and size(M,2) returns 5 (number of columns).
6. The Programming Environment
Creating Functions
Scilab has an open programming environment that enables users to build their own functions and libraries. It is done by using the built-in editor SciPad. To call the editor, type scipad() or editor(), or click Editor at the menu bar.
The file extensions used by scilab are .sce and .sci. To save a file, click for the menu File and choose Save. To load a file, choose Load under the same menu. To execute a file, type
exec('function_file_name');
in the command line, or click for load into Scilab under the menu Execute.
To begin writing a function, type:
function [out1, out2, ...] = name(in1, in2, ...)
where function is a keyword that indicates the start of a function, out1, out2 and in1, in2, ..., are variables that are outputs and inputs, respectively, of the function; the variables can be Boolean, numbers, matrices, etc., and name is the name of the function. Then we can enter the body of the function. At the end, type:
endfunction
to indicate the end of the function.
Comment lines begin with two slashes //. A sample function is given as below:
function [d] = distance(x, y) // this function computes the distance // between the origin and the point (x, y) d = sqrt(x^2 + y^2); endfunction
A Scilab function can call upon itself recursively. Here is an example.
Unlike Matlab, Scilab allows multiple function declaration (with different function names) within a single file. Also Scilab allows overloading (it is not recommended for beginners). Please refer to the chapter overloading in its help file for details.
Flow Control
A table of logical expressions is given below:
| == | equal |
| ~= | not equal |
| >= | greater than or equal to |
| <= | less than or equal to |
| > | greater than |
| < | less than |
| ~ | not |
If a logical expression is true, it returns a Boolean variable T (true), otherwise F (false).
The if statement
It has the basic structure:
if condition body end
The body will be executed only when the condition statement is true. Nested if statements have the structure:
if condition_1 body_1 elseif condition_2 body_2 elseif condition_3 body_3 elseif ... ... end
As an example, we make use of the fact that a Scilab function can call upon itself recursively to write a function which gives the n-th Fibonacci number in the sequence 0, 1, 1, 2, 3, 5, 8, 13, ... :
function [K] = fibonacci(n)
//function [K] = fibonacci(n)
//Gives the n-th term of the Fibonacci sequence 0,1,1,2,3,5,8,13,...
if n==1
K = 0;
elseif n==2
K = 1;
elseif n>2 & int(n)==n // check if n is an integer greater than 2
K = fibonacci(n-1) + fibonacci(n-2);
else
disp('error! -- input is not a positive integer');
end
endfunction
Note that this is not an efficient way to find the n-th Fibonacci number when n is large.
The for loop
It has the basic structure:
for variable = initial_value : step : final_value body end
The loop will be executed a fixed number of times specified by the number of elements in the array variable. A slightly modified version is:
str = 'abcdr'; s = ''; // an empty string for i = [1 2 5 1 3 1 4 1 2 5 1] s = s + part(str, i); end disp(s); s = abracadabra
The while loop
It has the basic structure:
while condition body end
The loop will go on as long as the condition statement is true. Here we give an example of the Euclidean Algorithm.
function [n1] = hcf(n1, n2) // n1 and n2 are positive integers if n2 > n1 tem = n2; n2 = n1; n1 = tem; // to ensure n1>=n2 end r = pmodulo(n1, n2); // remainder when n2 divides n1 n1 = n2; n2 = r; while r ~= 0 r = pmodulo(n1, n2); n1 = n2; n2 = r; end endfunction
The break and continue commands
They are for ending a loop and to immediately start the next iteration, respectively. For example:
// user has to input 10 numbers and for those which
// are integers are summed up, the program are
// prematurely once a negative number is entered.
// It is not well written but just to illustrate the
// use of the "break" and "continue" commands
result = 0;
for i = 1:10
tem = input('please input a number');
if tem < 0
break;
end
if tem ~= int(tem) //integral part
continue;
end
result = result + tem;
end
disp(result);
Some Programming Tips
The concept of Boolean vectors and matrices is important. The function find is useful too. It reports the indices of true Boolean vectors or matrices. For example:
M = [-1 2; 4 9]; M > 0 ans = F T T T M(M > 0)' ans = 4. 2. 9. end
In contrast,
find(M > 0) ans = 2. 3. 4. M(find(M>0))' ans = 4. 2. 9.
We remark that M(M>0) gives results quicker than M(find(M>0) because the find function is not necessary here.
It is important to distinguish & and and, | and or. The first one of each pair is entry-wise operation and the other one reports truth value based on all entries of a Boolean matrix.
M = [0 -2; 1 0]; M==0 | M==1 ans = T F T T and(M >= 0) // true iff all entries are true ans = F or(M == -2) // false iff all entries are false ans = T
Debugging
The most tedious work in programming is to debug. It can be done in two ways: either using Scilab's built-in debugger, or modifying the program so that it serves the same purpose as a debugger.
The Scilab debugger is similar to those debuggers in other programming languages and is simple to use. We present the second method to offer programmers greater flexibilities when debugging.
To insert breakpoints we use
pause
To end pause we use
abort
To set the output of the function we may use
return
To display variables we use
disp (variable)
For details please read the Scilab documentation.
7. Plotting Graphs
2D Graphs
The plot function has different forms, depending on the input arguments. If y is a vector, plot(y) produces a piecewise linear graph of the elements of y versus the index of the elements of y. When x and y are vectors of the same length, plot(x,y) produces a graph of y versus x. E.g., to plot the value of the sine function from zero to 2 pi:
t = (0:1/100:2) * %pi; y = sin(t); plot(t,y); |
 |
3D Surfaces
The command plot3d(x,y,z) plots 3D surfaces. Here x and y (x-axis and y-axis coordinates) are row vectors of sizes n1 and n2, and the coordinates must be monotone, and z is a matrix of size n1xn2 with z(i,j) being the value (height) of the surface at the point (x(i),y(j)).
// A simple plot of z = f(x,y) t=[0:0.3:2*%pi]'; z=sin(t)*cos(t'); plot3d(t,t,z) |
 |
8. Scilab versus Matlab
This section is based on some user comments found in the internet, thus not necessarily all true. It is intended to give readers a general image about their differences besides those in syntax.
- Matlab has a thorough documentation; the one in Scilab is brief.
- Matlab has a lot of optimization on computation, thus it is faster than Scilab.
- Matlab has a very powerful simulation component called Simulink.
- Scilab has Scicos that serves the same purpose but it is weaker.
- Matlab has a much better integration with other programming languages and programs such as C, C++ and Excel.
- The graphics component of Scilab is weak (has fewer functions).
- Most importantly, Scilab is FREE. It certainly outweighs its deficiencies. It is remarked that Scilab is more than enough for casual and educational uses.
9. References
- Scilab help file (its own documentation)
- Scilab for dummies
- Matlab primer
- A pratical introduction to Matlab
19 December 2006, revised 5 January 2009.