Scilab Based Global
Optimization Toolbox using the Bernstein Form
Interdisciplinary
Programme in Systems and Control Engineering,
Indian Institute of Technology
Powai, Mumbai (
------------------------------------------------------------------------------------------------------------
This project is funded by the open source software project of
National
I)
http://www.spoken-tutorial.org/NMEICT-Intro, II) http://scilab.in,
III) http://fossee.in
ABOUT TOOLBOX:
=============
This
is the version 1.0 (June 2010) of Global Optimization Toolbox using Bernstein
Form for SCILAB.
A Scilab based global optimization toolbox for
solving unconstrained polynomial optimization problems is developed. The
toolbox computes, to a user specified accuracy, the global minimum and all
global minimizers in a given initial search box.
The toolbox handles univariate as well as
multivariate type objective functions. The Bernstein polynomial approach serves
as the basis for the toolbox. Several optimization examples are provided along
with the toolbox to illustrate the approach and use of the software.
Keywords:
Bernstein polynomial, global optimization, multivariate polynomial,
unconstrained optimization
AUTHORS:
=======
Authors:
Mr. Dhiraj B. Magare, Mr. Bhagyesh V. Patil and
Prof. P. S. V. Nataraj
Indian Institute of Technology
E-mail: dhirajmagare@gmail.com,
bhagyesh@sc.iitb.ac.in
Installation:
=============
Before installing the toolbox, please read the
license file “license.txt”.
The installation of the toolbox is done in several
steps (More details can also be found in the file “overview.xml”)
1) If you do not yet have SCILAB, get it from ftp.inria.fr:INRIA/Scilab/
(Scilab 5.1.1 and above latest version)
2) For Installation, follow any one of the four
procedures (i.e. A or B or C or D).
-
A)
Recommend to follow this procedure for auto installation.
Run this command on console:
atomsInstall([‘Global_Optim_toolbox’,’1.0’])
* Command handles toolbox’s
installation automatically. (In Backend, it connects directly to ATOMS server.)
Suppose that <PATH> stands for the path of the
directory which contains the readme.txt file.
-
B) On Unix/Linux systems
* Administrator has to
execute only once, the following
instruction within Scilab:
exec
<PATH>/builder.sce
This operation
requires permission to write in
<PATH>/macros to generate *.bin, names and lib files.
* User should execute the
following instruction from within
Scilab:
exec <PATH>/loader.sce
before using
the toolbox. This statement can also be put
it in .scilab
startup file for automatic loading.
-
C) On Windows systems
* Administrator has to
execute only once, the following
instruction
from within Scilab:
exec
<PATH>\builder.sce
This operation
requires permission to write in
<PATH>\macros to generate *.bin, names and lib files
* User should execute the
following instruction from within
Scilab before
using the toolbox:
exec
<PATH>\loader.sce
The statement
can also put it in the .scilab startup file
for automatic
loading.
- D) Requires freeware ‘Inno Setup Script’ (ISS)
software, which
is available on following
weblink:
http://www.jrsoftware.org/isinfo.php
After installing ISS, run
exe file which is present in
‘Setup.zip’
Good
luck.
For version Scilab 5.2.2
Use following commands:
1)
pwd //Shows present working directory path
//Set Toolbox
directory path using File>Change Current Directory
2)
a=mopen(TMPDIR+'builder.sce','wt') // must give nonzero value
(like a=1)
3)
exec(‘builder.sce’)
//must have to generate ‘loader.sce’ under same directory
4)
exec(‘loader.sce’)
//must need to show you GUI asking for polynomial and more details read
in “overview.xml” file.
Contents
========
readme.txt
: first read this file
license.txt
: license information
loader.sce
: installation script
builder.sce
: script for building library
macros
: directory of Scilab functions
*.sci : source versions
*.bin : precompiled binary
versions (generated)
names : table of functions
(generated)
lib : Scilab
library binary save (generated)
help : directory for
help.
Etc : Containing
.start and .quit file for
default execution
demos : demo directory
doc : pdf document
directory
Download Link:
Or
Download ‘Inno Setup script (ISS)’ software from following link:
http://www.jrsoftware.org/isinfo.php
After installing ISS, run setup.
Toolbox Demonstration :
Once toolbox installed successfully, you
can see ‘Global Optimization Toolbox’ name under Scilab Demonstration option.
See List of Examples. Answers would be given
on Scilab command window:
Univariate Example :1
Ans: Global
Minimum zcap
- 2.
Univariate Example: 2
Ans: Global Minimum zcap
2.
Bivariate Example :
Six Hump Back Camel Function
Ans:
Two Variable Problem
Global Minimum zcap
- 1.03162845348967447
Multivariate Example :
Hairer Function
(Six Variables problem)
Ans:
Global Minimum zcap
-
1875.25
Univariate
Example with GUI Facility
Ans:
Global Minimum zcap
10.
VISITOR NUMBER