Re: [Help-glpk] Parametric studies in GMPL?

[Haroldo Santos]

Really great solution ! Appears as concise as AMPL and offers all the
flexibility of the Python language.


On Fri, Jul 9, 2010 at 5:23 PM, Xypron <address@hidden> wrote:
> Hello Haroldo,
>
> in GMPL it is only possible to solve a single MIP or LP problem.
>
> Using the GLPK-API it is possible to implement heuristics like the one you
> describe.
>
> Another approach is to use a data base for storage of partial solutions and
> a scripting language to call glpsol for different problems.
>
> Commercial linear solvers provide there own scripting languages, e.g.
> ILOG OPL Studio, Xpress Mosel
>
> If you want to stick with GLPK consider using Python with package PyMathProg
> http://pymprog.sourceforge.net/
>
> An interesting example how PyMathProg can be used for implementing a
> heuristic is
> http://pymprog.sourceforge.net/subtour.html
>
> Best regards
>
> Xypron
>
> Haroldo Santos wrote:
>>
>> I'm interested in something similar, I'm not quite sure if it is
>> possible to do in GMPL, but it would certainly be nice:
>>
>> Consider that you have a problem to solve, which is too hard
>> considering MIP capabilities.
>>
>> I would like to do, in GMPL, some (heuristic) like:
>>
>> Split problem P into P1, P2 ...  Pn
>> Solve P1
>> Solve P2
>> ....
>> Solve Pn
>> merge solution of P1  ... Pn to provide a solution to P
>>
>> Is there any way to do it ?
>>
>> Thanks,
>>
>> Haroldo
>>
>>
>>
>>
>
>



-- 
=============================================================
Haroldo Gambini Santos
Computing Department - Universidade Federal de Ouro Preto - UFOP
email: haroldo [at ] iceb.ufop.br
home/research page: http://www.iceb.ufop.br/decom/prof/haroldo/

"Computer science is no more about computers than astronomy
is about telescopes." Edsger Dijkstra