[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-glpk] [Fwd: --tmlim 3600 didn't work for large scale MILP probl
From: |
Andrew Makhorin |
Subject: |
Re: [Bug-glpk] [Fwd: --tmlim 3600 didn't work for large scale MILP problem.] |
Date: |
Mon, 13 May 2013 17:35:37 +0400 |
> Solving a large scale MILP problem I set time limit to one hour. Four
> hours later the model was still running...
>
>
> I think, as it didnĀ“t find any solution and time check procedure may
> happen between these process.
>
>
> I attach the log file, after forcing abrupt termination.
>
Thank you for your report.
The problem is that the time limit is checked in the b&b solver,
however, the dual simplex solver falls into infinite loop and never
returns to the b&b solver.
Most likely that your instance is hard for glpk due to its size,
besides, it looks badly scaled. You may try to disable automatic scaling
with --noscale option and also disable gomory cuts.
Andrew Makhorin