MIP bound is greater than reported incumbent solution and reported solution is suboptimal
AnsweredHi,
when solving a MIP using ConcurrentMIP (as well as some other non-default parameter settings, see linked log-file), Gurobi reports multiple weird things:
- The reported bound is greater than the incumbent.
- The log shows that for several iterations before termination, the best bound was already greater than the incumbent (yet this is a different bound than that reported at the end).
- The reported best objective does not match the one given in the log in the line "No other solutions better than ..."
- The solution in the log line "No other solutions better than ..." is clearly suboptimal.
The MPS file and log are here: https://rwth-aachen.sciebo.de/s/kGG6AqUeYAz2ZVv
I am running Gurobi 9.5.0 (upgrading to 9.5.2 is not easy because I am not in control of software updates on the machine), so I'm not 100% sure if this may be a bug fixed in one of the two releases since then.
Cheers,
Adrian
-
Related to Number of Threads affects MIP Solution
0 -
Thanks for reporting the issue! This appears to be the same bug that you reported in the post Gwyneth linked. Upgrading to Gurobi 9.5.2 should resolve the issue, though I understand that's not necessarily easy for you.
0
Please sign in to leave a comment.
Comments
2 comments