Re: octave-glpk MIPGAP

[Marcelo Pinto]

Hi Jordi,

Thanks for reply.

I'm not an expert in C++, but I have modified the __glpk__.cc source
file and added the mipgap functionality. Now the optimizer stop the
execution when the tolerance specified in the new "param.mipgap" is
reached.

However, as the code was not prepared to have this functionality, the
vector with the variables values that minimize or maximizer the
objective function is returning with "N/A" values.

I haven't figured out how can I fix this yet. If you have some tip
about what can I do, I would be very thankful.

Tks,
Marcelo.

2011/3/27 Jordi Gutiérrez Hermoso <address@hidden>:
> On 26 March 2011 21:08, Marcelo Pinto <address@hidden> wrote:
>> I'm trying to convince the people here in our firm to migrate from
>> Matlab to Octave. However, the MIPGAP feature of GLPK is very
>> important in our applications, and this functionality isn't
>> implemented in octave's binding for glpk.
>>
>> Someone can help me how can I do to fix this problem? If I have to
>> implement this feature, what source codes should I have to modify?
>
> The GLPK code is spread across two source files (btw, in English,
> source code is a mass noun, like help and information, there are no
> helps, no informations and no source codes):
>
>     
> http://hg.savannah.gnu.org/hgweb/octave/file/tip/scripts/optimization/glpk.m
>     
> http://hg.savannah.gnu.org/hgweb/octave/file/tip/src/DLD-FUNCTIONS/__glpk__.cc
>
> As you can see, the argument parsing is handled in glpk.m but the
> actual work is done in __glpk__.cc. I am not familiar with GLPK to
> tell you how difficult it would be to add more functionality to the
> code, but I can help you with the infrastructure of writing and
> committing a patch if you would like.
>
> HTH,
> - Jordi G. H.
>