[Coin-discuss] Problem with a Branch and Price and dual variables

[Matthew Galati]

Given that the dual ray (or proof of infeasibility) comes from the basis 
(and, it is possible that more than one row provides a proof), there is 
no way to guarantee that two different implementations of simplex will 
return the same infeasibility ray.

Can you explain more about the problems you are seeing using CLP and/or CPX?

Matt



> hi,
> 
> I am trying to implement a column gerneration algorithm,
> and had some problems with dual rays to restore feasibilty after
> branching.
> So I wanted to check if cplex and CLP a returning the same values, in
> order to determine where my problem is. I turn presolve off for cplex,
> and give hints to CLP to turn presolve and DualSimplex off.
> The problem is that after the first iteration I allready get different
> dual variables back so the prices for column generation are different.
> I allready looked at the .mps files of the solvers: they are the same.
> 
> I would be very gratefull if somone could give me a hint how to solve
> this.
> 
> thank you in advance!
> best wishes
> jakob
> 
> 
> --
> Dipl.-Ing. Jakob Puchinger
> Algorithms and Data Structures Group
> Institute of Computer Graphics and Algorithms (E186)
> Vienna University of Technology
> Favoritenstraße 9-11, A-1040 Vienna, Austria
> email: japu at ads.tuwien.ac.at
> phone: +43(1)58801/18627
> http://www.ads.tuwien.ac.at/
> --
> 
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss

-- 
Matthew Galati
ISE Lehigh University
IBM Service Parts Solutions
610.758.4042 (Office)
610.882.0779 (Home)
magh at lehigh.edu, magal11 at us.ibm.com
http://sagan.ie.lehigh.edu/mgalati/