[SCIP] strong branching and cuts

[James Cussens]

As I understand it, when running strong branching on a candidate branching variable, the 'down-branch' and 'up-branch' LPs are solved and we can inspect the associated dual bounds to measure how good a choice the candidate would be to branch on.

I can't work out whether cuts are added to these two LPs and, if so, how to prevent them being added. Could someone enlighten me?

Evidently, if the down and up LPs undergo some separation rounds then we get more useful dual bounds and thus a better idea of the quality of the variable as something to branch on. However, if those separation rounds are expensive, it might be better to skip them.

James

James Cussens
Room MVB 3.26
School of Computer Science, University of Bristol
Phone: +44 (0)117 455 8723
https://jcussens.github.io/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20250305/9d96e6c7/attachment.html>