[SciPy-user] citing numpy, scipy openopt

[Emanuele Olivetti]

dmitrey wrote:
> ...
> it would be nice to know where openopt is used, both for me and other OO 
> users. IIRC TOMOPT informed his users that their TOMLAB is used at 
> international space station. Mb someday OO will be used so commonly as 
> well?:)
>   

Surely you noticed that there is the need of having a free software /
opensource alternative to tomopt/tomlab. Since python+numpy+scipy
are gaining popularity you can only expect a larger user base, as
time goes on.

>> In the meanwhile could you send a basic reference about the ralg
>> algorithm? 
>>     
> There are lots of ralg modifications: ralg-5 (initial version, that is 
> implemented currently in OO), ralg-4 (more advanced implementation, it 
> requires only 4n^2 multiplications vs 5n^2 in ralg-5). As for Fortran 
> version, it is capable of handling nVars ~= 10^4...10^5, while ordinary 
> ralg only ~1000. Also, it consumes much less memory: m*nVars vs nVars^2 
> (m<<nVars, for nVars ~1e4...1e5 in a factor of 100 or so). However, it's 
> not pure r-alg, there are some modifications of course, as well as 
> different ways of handling constraints (pure r-alg is unconstrained).
>
> You could try google "N.Z.Shor" or "Naum Z. Shor" for references.
>   

>From wikipedia I've found this document "Congratulations to Naum
Shor on his 65th birthday" from Journal of Global Optimization. It has
a nice list of references which is presented in the text:
http://www.springerlink.com/content/j263467v12w40172/fulltext.pdf
(I can access it, but I don't know if it is free access)

I do believe you should put similar references, contextually, in the
documentation of your OpenOpt library. As David said, it is common
practice to mention such references to algorithms (and possibly software)
when writing scientific articles. It should be definitely relevant for your
department (so directly or indirectly for you).


Emanuele


P.S.: I'm sorry for the inappropriate question about asking Shor.