help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: recursion and fsolve


From: Bart Vandewoestyne
Subject: Re: recursion and fsolve
Date: Fri, 17 Sep 2004 09:51:32 -0500
User-agent: Mutt/1.3.28i

On Fri, Sep 17, 2004 at 06:46:17AM -0700, Paul Laub wrote:
> Dear all, 
> 
> All of this talk about different methods of generating
> pseudo random deviates raises the question: 
> 
>     Are there m-files publicly available for testing these
>     generators for serial correlation and other problems?
> 
> With such tests, one might be able to more
> intelligently evaluate the tradeoff between quality
> of pseudo random deviates and speed, portability of
> code, etc. 
> 
> If I remember right, Donald Knuth discusses such tests
> at length in one of his Art of Computer Programming
> volumes. Someone must have implemented them.  

There are two things i would like to mention:

1) First of all, i think the Octave community has to decide what she
wants: generating the same pseudo-random numbers as Matlab (= using the
same random number generator, with its possible advantages/drawbacks) or
using another random number generator which might have better properties
than the one currently built-in in Matlab.
I don't know if it is a possibility to let the rand function have an
optional parameter such as 'GeneratorType' or something like that...
then the user could choose which RNG he uses.  The default when not
specifying a GeneratorType could be the same as Matlab.  Could this be a
possibility?

2) Concerning the tests for random number generators: i don't know of
any m-files that implement such tests, but maybe these links can be
helpfull in learning about such tests and implementing them in m-files:

http://random.mat.sbg.ac.at/
http://stat.fsu.edu/~geo/
http://csrc.nist.gov/rng/
http://www.taygeta.com/random.html


Hmmm... while I'm writing this, couldn't this be a nice subject for a
Master's thesis of somebody doing compuer science?  If the Octave people
feel the need for a RNG testsuite in m-files, i might talk about this
to my promotor (I'm a PhD student myself) and ask him if I could write
this out as a masters thesis proposal...  If i get a 'yes' on this, then
the student would start somewhere in August 2005 and finish somewher in
June 2006...

Regards,
Bart

-- 
!!!!!!!!!!!!!!!!!!! email change !!!!!!!!!!!!!!!!!!!!!!!
My email address is now Bart.Vandewoestyne AT telenet.be
        Please update your addressbook!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------



reply via email to

[Prev in Thread] Current Thread [Next in Thread]