[Cbc] How to stop execution of CBC. Is it possible to see current result?
John Forrest
john.forrest at fastercoin.com
Tue Mar 22 15:45:12 EDT 2016
Jonatan,
It has found a solution of 4027.85. Use allowablegap or ratiogap to
stop when gap is small.
John Forrest
On 21/03/16 08:31, Jonatan Knud Lauritsen wrote:
> I am running my programming task with command "cbc model_cplex.lp
> solve solu solution.txt" and now CBC is already working for almost
> hour and giving messages like:
> Cbc0010I After 75400 nodes, 16912 on tree, 4027.85 best solution, best
> possible
> 4026.84 (874.05 seconds)
> Cbc0010I After 75500 nodes, 16914 on tree, 4027.85 best solution, best
> possible
> 4026.84 (876.98 seconds)
> Cbc0010I After 75600 nodes, 16914 on tree, 4027.85 best solution, best
> possible
> 4026.84 (880.22 seconds)
> Cbc0010I After 75700 nodes, 16913 on tree, 4027.85 best solution, best
> possible
> 4026.84 (882.89 seconds)
> Cbc0010I After 75800 nodes, 16913 on tree, 4027.85 best solution, best
> possible
> 4026.84 (885.77 seconds)
> Cbc0010I After 75900 nodes, 16913 on tree, 4027.85 best solution, best
> possible
> 4026.84 (888.34 seconds)
> Cbc0010I After 76000 nodes, 16913 on tree, 4027.85 best solution, best
> possible
> 4026.84 (891.18 seconds)
> Cbc0010I After 76100 nodes, 16906 on tree, 4027.85 best solution, best
> possible
> 4026.84 (894.64 seconds)
> Cbc0010I After 76200 nodes, 16894 on tree, 4027.85 best solution, best
> possible
> 4026.84 (897.10 seconds)
> Cbc0010I After 76300 nodes, 16885 on tree, 4027.85 best solution, best
> possible
> 4026.84 (899.48 seconds)
> Cbc0010I After 76400 nodes, 16900 on tree, 4027.85 best solution, best
> possible
> 4026.84 (902.47 seconds)
> Cbc0010I After 76500 nodes, 16902 on tree, 4027.85 best solution, best
> possible
> 4026.84 (904.88 seconds)
>
> What does this mean? As I understand, them my problem is feasible but
> what CBC is trying to find so long? Is it trying to find first
> feasible integer solution or CBC has already found some solution but
> now is trying to find better solution. I would be satisfied with
> suboptimal solution if it takes less time to find. From smaller
> problems I know that first feasible solution usually is quite good,
> the next best solutions required lot of time to be found and the
> improvements are small.
>
> Jonatan
>
>
> _______________________________________________
> Cbc mailing list
> Cbc at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cbc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20160322/8f9a208e/attachment.html>
More information about the Cbc
mailing list