[Help-glpk] Dantzig-Wolfe Decomposition and GLPK

[Joey Rios]

Hello,

I was wondering if anyone had any experience trying to implement some form of Dantzig-Wolfe decomposition using GLPK.  If so, I'd love any info/advice on how you got started.  I am reading up on the algorithm now, so assuming I get a handle on that, any insight on actual implementation would be appreciated.  I understand there is a way to handle the problem using the AMPL or GAMS modeling language, but I think some of the necessary constructs are not implemented in MathProg, which I have been using.  I am getting more comfortable with the GLPK API, so implementing the algorithm in C would be fine (and probably the only way?).  Or can something like this example be ported to MathProg:  http://www.gams.com/modlib/libhtml/danwolfe.htm ?

Anyway, just looking for some tips from anyone who may have been down this path before.

I can clarify or expand on any of my questions if that is helpful.  Thanks in advance.

Joey

---

Need to know the score, the latest news, or you need your Hotmail®-get your "fix". Check it out.