Re: [Help-glpk] loop for for mathprog

[Andrew Makhorin]

Marc,

Thank you for interest in glpk and its modeling language.


> When I read the documentation, it appears that the AMPL like
> abilities in GLPK do not include the for-loop control structures. In
> particular, there is very little you can do after the “solve”
> command (basically you can use “display” and “print”, but there is
> no facility for examining the solution and rerunning the linear or
> integer program).
>  
> In “real” applications of linear and integer programming, I almost
> always need some post-solution control due to either:
>  
> (1) Decomposition techniques (such as column generation)
>
> (2) Having the linear program only able to approximate the
> real-world problem. In this case, the solutions need to be analyzed
> and the linear program needs to be adjusted several times before an
> acceptable solution that fits the “real-world” constraints is
> obtained.
>  
> So my number #1 wish for Andrew is to add in the same type (or
> better) post-solution control that AMPL has.

There is a dilemma, because I do not think that mixing modeling
and "programming" capabilities within model description (as in
AMPL) is a good idea; on the other hand I would like to keep
compatibility between GNU MathProg and AMPL. Probably, there must be
another level, something like a script, that allows performing all
operations needed for reoptimization and post-optimal processing.

In principle, optimization results could be written to a file in the
form of data section and used in other (or the same) model to perform
reoptimization or post-optimal processing. This is not an elegenat way,
however, it is in the spirit of "programming". :)


Andrew Makhorin