I think we really need to have a program fragment. I just discussed with Guido on the phone, from our understanding getti9ng the same id twice is impossible ;-) As it is not we really have to see what you do.
Cheers Christian -- Christian Schulte, http://www.imit.kth.se/~schulte/ -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sébastien Mouthuy Sent: Tuesday, May 29, 2007 12:24 PM To: Guido Tack Cc: [EMAIL PROTECTED] Subject: Re: [gecode-users] order in calls to status and commit... Le mardi 29 mai 2007 à 13:07 +0200, Guido Tack a écrit : > Sébastien Mouthuy wrote: > > > I have one main branching, let's say BrSF. In the commit function of > > BrSF, I sometimes post another branching, say BrCl. > > > > I printed id's of these branching and it happens that an instance of > > BrSF has the same id as BrCl. I'm not using recomputation: c_d = 1, > > a_d > > = 10000. > > This is definitely a problem. The same id for different branchings > may only occur on different branches of the search tree, but not on > the same branch! Are the BrSF also created dynamically? No. > Do you, by > any chance, create a branching in the commit-function you use for > branch-and-bound? > I don't understand this question. I am using a "constrain" function in my Space to use Branch-and-Bound. What do mean by that ? Thanks, sebastien > Anyway, I'm sure the problem with the ids is also the reason for the > original problem. Please try to find out why two different branchings > get the same id. > > Cheers, > Guido > _______________________________________________ Gecode users mailing list [EMAIL PROTECTED] https://www.gecode.org/mailman/listinfo/gecode-users _______________________________________________ Gecode users mailing list [EMAIL PROTECTED] https://www.gecode.org/mailman/listinfo/gecode-users
