[Scip] Tweaking SCIP parameters for speed-up

[Sergey Smirnov]

Hello,

I am writing a tool which can optimize SCIP's parameters to make a given set of problems solve faster. After testing with several problem sets it seems that the tool is quite promising as it produces a configuration file which provides at least 2x speedup in comparison to SCIP's default parameters. The parameters it finds are quite different for different kinds of problems.

Does anyone has a problem to speed up? I can help and I am eager to obtain more data about how my approach performs.

Here is a brief description of how the tool works. Firstly it estimates how long it takes to solve single problems with default parameters. Then we do a single run (for every problem) for every special value (e.g. -1 or 0) of every parameter and two runs for ranged parameters (bigger and lower than default values). That is about 3000 runs for every problem in a test set so I use a cluster of virtual machines in a cloud for the computations. After this step the tool compares the impact which parameters and their values make on the solving time. Finally the tool takes four parameters (and their values) with the biggest impact and finds their best combination. This step is repeated for a fey times and we get a settings file which speeds up computations noticeably for the given set of problems.

I have not upload the source code anywhere yet because it is not well documented. However if anyone is interested I can put it on GitHub.

Best regards,
Sergey Smirnov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20140710/96dbe1c7/attachment.html>