[SCIP] solving with a SOC constraint

[James Cussens]

Hi all,

I have a pricing algorithm which solves a MIQP which has a single SOC constraint, some SOS-1 constraints and a number of linear constraints.

I worked out some lower and upper bounds for some of my continuous variables (with a view to perhaps replacing the SOS-1 constraints by some bigM ones).
I have observed that merely putting these bounds on the variables in question, rather than having bounds of -infinity and infinity (and doing nothing else) leads to a big drop in performance, solving is slower.

I am not using an NLP solver.

Any ideas why this might happen? Possibly adding the bounds rules out some sub-optimal solutions which are helping somehow.

The bounds are rather loose, but I did not expect them to be deleterious.

James


James Cussens
Dept of Computer Science, University of Bristol
https://jcussens.github.io/
Funded PhDs available in Bristol in the following areas: Data Science<http://www.bristol.ac.uk/cdt/compass/>, Interactive AI<http://www.bristol.ac.uk/cdt/interactive-ai/>, Cyber Security<http://www.bristol.ac.uk/cdt/cyber-security/> or Digital Health<http://www.bristol.ac.uk/cdt/digital-health/>.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20210218/a653b3c0/attachment.html>