[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-glpk] FPUMP heuristic pass bug GLPK4.60
From: |
Andrew Makhorin |
Subject: |
Re: [Bug-glpk] FPUMP heuristic pass bug GLPK4.60 |
Date: |
Wed, 06 Apr 2016 13:42:09 +0300 |
> I think FPUMP Heuristic pass routine has a bug in GLPK4.60
>
> When Fpump finds a solution, it loop back again to the beginning of
> pass instead of continuing from the current npass.
>
> See the log below.
>
> Noli
> ~~~~~~~~~
> .
> .
> * 10500: obj = 1.610401485e+05 inf = 8.027e-12 (5255) 3
> * 11000: obj = 1.611132718e+05 inf = 7.347e-12 (413) 3
> * 11279: obj = 1.611244848e+05 inf = 1.124e-10 (0) 2
> OPTIMAL LP SOLUTION FOUND
> Integer optimization begins...
> WARNING: LONG-STEP DUAL SIMPLEX WILL BE USED
> + 11279: mip = not found yet <= +inf (1; 0)
> Applying FPUMP heuristic...
> Pass 1
> Pass 2
> Solution found by heuristic: 147859.9
> Pass 1
> Pass 2
> Pass 3
> Pass 4
> Pass 5
> + 11279: mip = 1.478599000e+05 <= 1.611235344e+05 9.0% (2; 0)
> Time used: 86.3 secs. Memory used: 107.8 Mb.
> + 11279: mip = 1.478599000e+05 <= 1.611235344e+05 9.0% (2; 0)
> RELATIVE MIP GAP TOLERANCE REACHED; SEARCH TERMINATED
>
No, it is not a bug (it is a feature). If FPUMP has found a solution, it
repeats the search to find a better solution until 5 successive attempts
fail.