Re: gnumeric solver

[Jean Bréfort <jean brefort normalesup org>]

Looks interesting. I also would like a non linear solver in gnumeric.
Gnumeric code is written in C, not C++, but I think a plugin written in
C++ would be acceptable.

Best regards,
Jean

Le dimanche 17 août 2008 à 21:08 +0200, Ali Baharev a écrit :
> Dear Developers,
>
> I am a PhD student at  Budapest University of Technology and
> Economics, i have an MSc in chemical engineering.  Please note that i
> have never learned C++ programming officially.
> Currently i am working on a non-linear system solver written in C++
> based on a new branch of interval arithmetic called affine arithmetic
> or GI form. What i have implemented is the GI form discussed in the
> paper below:
>
> http://www.solver.com/IntervalSolverRC.pdf
>
> Slightly more details and papers on my website:
>
> http://reliablecomputing.eu/
>
> I would like to ask you for your help. The code i have done so far
> make little if any use to others, however if i could integrate it into
> Gnumeric as a solver it would be useful for many people - i hope.
>
> There are several problems. I use abstract data types, all operators
> and functions are overloaded. The algorithm finds all solutions, it is
> not clear to me how to manage that. My solver depends on GLPK.
>
> Could you please help me how to integrate an interval solver?
>
> Another "solution" could be to access the cell contents as symbolic
> strings in my C++ code. If i could get the cell entry as F1=C1+D1
> where in turn C1=A1*B1, this would probably work for me.
>
> Thank you for your help,
>
> Ali Baharev
> _______________________________________________
> gnumeric-list mailing list
> gnumeric-list gnome org
> http://mail.gnome.org/mailman/listinfo/gnumeric-list