[SCIP] implementing a branching strategy

Gerald Gamrath gamrath at zib.de
Tue Mar 15 14:23:06 CET 2016


Hi James,

if you just want to do this with relpscost, you could call 
SCIPexecRelpscostBranching() in the ENFOLP callback of your constraint 
handler and only pass the subset of variables as branching candidates.

Best,
Gerald

On 15.03.2016 12:18, James Cussens wrote:
> Dear SCIPers,
>
> I would like to implement a branching strategy where at each node only 
> a subset of variables are considered as branching candidates. Which 
> subset this is will depend on the node in question (specifically which 
> variables are already fixed by the time we get to that node).
>
> It seems to me that a potentially easy way of doing this is to update 
> the branching priorities of variables before the branching decision is 
> made (I want the current branching rule (eg relpscost) to choose which 
> of my candidates to branch on.)
>
> But I'm not sure
> 1) if this is a reasonable way of implementing what I want to happen, and
> 2) if it is, where in my code to effect the branching priority updating.
>
> I have a constraint handler (it's a "global" one - there is always a 
> constraint which is not satisfied until the entire problem is solved). 
> One option would seem to be to update the branching priorities in its 
> ENFOLP callback and declare SCIP_INFEASIBLE. But this seems a bit of hack.
>
> Another possibility is to write an event handler, so that branching 
> priorities get updated when other variables get fixed.
>
> Any advice / ideas?
>
> James
>
> -- 
> James Cussens
> Dept of Computer Science &
> York Centre for Complex Systems Analysis
> Room 326, The Hub, Deramore Lane            Tel    +44 (0)1904 325371
> University of York  Fax  +44 (0)1904 500159
> York YO10 5GE, UK http://www.cs.york.ac.uk/~jc 
> <http://www.cs.york.ac.uk/%7Ejc>
> http://www.york.ac.uk/docs/disclaimer/email.htm
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20160315/f443b2fd/attachment.html>


More information about the Scip mailing list