[SCIP] Question about node numbering

[Marc Pfetsch]

Hi Franek,

SCIPgetSolNodenum() returns what sol.c:solStamp() enters, namely 
stat->nnodes, which are the number of nodes processed in the current 
run. SCIPnodeGetNumber() returns stat->ncreatednodesrun at the time of 
creation of the node. The documentation seems to be misleading here.

Thus, these numbers can deviate, since several child nodes might have 
been added before they are processed. Therefore some differences are 
expected, but what you describe seems a bit strange. But keep in mind 
that there might be several runs.

Best

Marc


On 30/06/2025 09:05, Franek Stark wrote:
> Hi,
> 
> I am using event handlers to track the B&B process for later ML 
> applications.
> Thereby, I encounter the following issue:
> 
> When a  SCIP_EVENTTYPE_SOLEVENT is triggered, I extract the solution 
> with  SCIPeventGetSol(). Now, the node number associated with this 
> solution (I get it via  SCIPgetSolNodenum() ), does not refer to the 
> node number of the current node being processed. It sometimes even 
> refers to Node numbers, from nodes in the future or from nodes that have 
> been pruned as infeasible a long time ago in the solving process.
> 
> Does SCIPgetSolNodenum() refer to a different node numbering than 
> SCIPnodeGetNumber() ?
> 
> Thanks for any help on this.
> Best regards,
> 
> Franek Stark
> 
> 
> 
> ---
> 
> 
> *Franek Stark, M.Sc.*
> 
> *Robotics Researcher*
> 
> *Robotics Innovation Center*
> 
> Deutsches Forschungszentrum für Künstliche Intelligenz GmbH (DFKI)
> 
> Robotics Innovation Center
> 
> Robert-Hooke-Str. 1
> 
> 28359 Bremen
> 
> Deutschland / Germany
> 
> Telefon: +49 421 178 45 4112
> 
> *Weitere Informationen: *
> 
> https://www.dfki.de <https://www.dfki.de/>
> 
> Firmensitz:
> 
> Trippstadter Straße 122, D-67663 Kaiserslautern
> 
> Geschäftsführung:
> 
> Prof. Dr. Antonio Krüger (Vorsitzender)
> 
> Helmut Ditzer
> 
> Vorsitzender des Aufsichtsrats:
> 
> Dr. Ferri Abolhassan
> 
> Amtsgericht Kaiserslautern, HRB 2313
> 
> 
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip