Wiki news

[Michaël Baudin]

Hi,

We recently updated several wiki pages which might be interesting for 
Scilab users.


Scilab can be used for thermodynamics calculations. For example, 
chemical equilibrium may result in nonlinear systems of equations, which 
can be solved by the Scilab function fsolve. An example of this is given 
in a paper by Assis, Hori, Avila-Neto and Franco, four researchers from 
the School of Chemical Engineering of the Federal University of 
Uberlandia in Brazil. They have used Scilab to conduct a comparative 
thermodynamic analysis of methane reforming reactions and to asses the 
influence of key operational variables on chemical equilibrium. On the page:

http://wiki.scilab.org/ThermodynamicsInScilab

we present some toolboxes which are providing thermodynamics features 
for Scilab. We review two toolboxes in this field: AmsterChem's 
toolboxes for Scilab, based on the CAPE-OPEN standard and Excel 
Engineering's toolbox for water and steam computations in Scilab.


During the year 2011, we have seen several new computational toolboxes 
or scripts, including:
  * 16th of February 2011: New Script : AG_fluid_dynamics
  * 3rd of March 2011: New Module: Wavelab
  * 8th of March 2011: New Module: Particle Swarm Optimization Toolbox
  * 15th of March 2011: New Module: FreeASM
  * 23rd of March 2011: New Script: Crystal Structure Animation
  * 6th of June 2011: New Module: Linear Algebra
  * 8th of June 2011: New Module: Mathieu functions
  * 20th of July 2011: New Module: NIST Datasets

We updated the "New Scientific Features in 2011" wiki page:

http://wiki.scilab.org/New%20Scientific%20Features%20in%202011

with reviews of these toolboxes.


The Professor Sakurai, from the University of Tsukuba has presented a 
part of his work with Scilab at the ScilabTec 2011: "Scilab for 
Mathematical Education and High Performance Computing". The slides from 
this conference are available at:

http://wiki.scilab.org/Tutorials#Slides_from_ScilabTec_2011

In his presentation, Prof. Sakurai discusses the use of Scilab to solve 
very large sparse systems of linear equations (requiring more than 10GB 
memory) and emphasizes the need for Scilab 6. This was the opportunity 
to make an overview of Scilab features for sparse systems of equations. 
In the page :

http://wiki.scilab.org/Solving%20Poisson%20PDE%20with%20Sparse%20Matrices

we present the resolution of the Poisson Partial Differential Equation 
in Scilab with sparse matrices. We show that Scilab 5 can solve in a few 
seconds sparse linear systems of equations with as many as 250 000 
unknowns because Scilab only store nonzero entries of the matrix. The 
computations are based on the Scibench module, a toolbox which provides 
a collection of benchmarks for Scilab.


Do the optimization solvers in Scilab perform well, compared to other 
solvers available as open source softwares ? In the page:

http://wiki.scilab.org/PerformancesScilabMoreGarbowHillstromBenchmark

we report benchmarks of unconstrained nonlinear optimization in Scilab. 
We analyze the performance of various optimization solvers on the Moré, 
Garbow and Hillstrom collection of test problems for unconstrained 
optimization and nonlinear least squares problems. More precisely, we 
tested the fminsearch, optim, lsqrsolve and leastsq nonlinear 
optimization solvers.


The Mandelbrot set is one of most famous fractals, although Julia sets 
are often presented as well. Given the iterative process used in the 
creation of the Mandelbrot set, it seems that there might be a 
performance issue, so that Scilab may be able to create only very small  
fractals of this type. In the page:

http://wiki.scilab.org/MandelbrotSet-NaiveVsVectorized

we present how to efficiently create a plot of the Mandelbrot set in 
Scilab. We emphasize the performance difference between a naive 
algorithm and a vectorized algorithm. We show how Scilab's complex 
datatype can be used to simplify the algorithm and increase the speed.


Best regards,

Michaël Baudin


-- 
Michaël Baudin
Ingénieur de développement
michael.baudin at scilab.org
-------------------------
Consortium Scilab - Digiteo
Domaine de Voluceau - Rocquencourt
B.P. 105 - 78153 Le Chesnay Cedex
Tel. : 01 39 63 56 87 - Fax : 01 39 63 55 94