[Coin-discuss] Suggestions on how to go about tuning?

[John J Forrest]

Philip,

If you can send me the model I will look at it and try and document the 
way I tuned it.  I might include such an example in my tutorial at DIMACS 
workshop if people think that a good idea.

John Forrest



hpwalton at comcast.net 
Sent by: coin-discuss-bounces at list.coin-or.org
06/01/2006 12:29 AM
Please respond to
Discussions about open source software for Operations Research 
<coin-discuss at list.coin-or.org>


To
coin-discuss at list.coin-or.org
cc

Subject
[Coin-discuss] Suggestions on how to go about tuning?






Is there doc anywhere on suggestions for tuning CBC/Clp? 
 
I'm having a performance issue with a specific instance of data.  Other 
instances of the same data run very quickly...(a few seconds), and this is 
running in terms of hours.
 
I'm thinking CBC just gets off on the wrong foot.  I'm able to solve it 
with another open source solver very rapidly (in the "few minutes" range). 
 
 
I was hoping I could learn enough on my own, but at this point I'm not 
really sure the most effective way to go about it.  I'm not looking for 
someone to do the thinkin' for me..more lookin' for some guidance on CBC 
specifically.
 
It is (roughly) a Hamiltonian Cycle problem...I've indicated some SOS type 
1 constraints and that helped get an answer fast, but the answer isn't all 
that good.
 
It's a minimization problem, the optimal solution is 2.1, the first 
integer answer CBC finds is 3.6, and the "lower bound" is 1.  the 
cost/weights are non-negative small decimals (ranging from zero to 2.5).
 
The lower bound doesn't start coming up for quite a while...so it spends a 
lot of time exploring nodes which are not fruitful.
 
Any suggestions are welcome.  I've seen at least one other request in the 
archives for tuning paramters...
 
Thanks,
Philip_______________________________________________
Coin-discuss mailing list
Coin-discuss at list.coin-or.org
http://list.coin-or.org/mailman/listinfo/coin-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/coin-discuss/attachments/20060601/27e8085f/attachment.html>