[Cbc] Cbc callable vs Cbc stand-alone

[Andrea Manno]

Dear all,


I am using the callable version of CBC as the MILP solver unit included in a C++ code. So what I have done is nothing but including all the Cbc, Osi and Cgl header file needed and linking all the necessary libraries in order to use the branchAnbBound() method of the CbcModel class. However, I noticed a deterioring of performance (in terms of CPU-time) compared to the stand-alone executable version of Cbc used as AMPL solver. I managed to eliminate this gap of performance for relatively small problems, by adding some Cbc heuristics and cuts, but the problem persists for bigger problems. I would like to know if there is a way to tune all the preprocess, cuts, heuristics and other options in order to replicate the same performance of the stand-alone version of Cbc.


Thank you so much,


Andrea Manno

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20180221/f1d493d8/attachment-0001.html>