[SCIP] Interpreting output
Marc Pfetsch
pfetsch at mathematik.tu-darmstadt.de
Tue Aug 3 15:47:43 CEST 2021
Hi Marco,
you can type "display display" in the interactive shell to get a
description of the different columns.
Each line of output, except for the first that corresponds to the output
of a heuristic, corresponds to one LP solve. In this case, they seem to
come from the fact that one cut is added after each solve (see "rows"
column).
Best
Marc
On 03/08/2021 14.55, Marco Correia wrote:
> Hi,
>
> I'm currently solving a problem where I get this output:
>
> time | node | left |LP iter|LP it/n|mem/heur|mdpt |vars |cons |rows
> |cuts |sepa|confs|strbr| dualbound | primalbound | gap | compl.
> t 0.7s| 1 | 0 | 0 | - | trivial| 0 |1580 |8299 | 0 |
> 0 | 0 | 0 | 0 | 2.443041e+00 | 0.000000e+00 | Inf | unknown
> 0.8s| 1 | 0 | 1326 | - | 47M | 0 |1580 |8300 |8298 |
> 0 | 0 | 1 | 0 | 1.708672e-02 | 0.000000e+00 | Inf | unknown
> 0.8s| 1 | 0 | 1326 | - | 47M | 0 |1580 |8300 |8295 |
> 0 | 0 | 1 | 0 | 1.708672e-02 | 0.000000e+00 | Inf | unknown
> 0.8s| 1 | 0 | 1328 | - | 49M | 0 |1580 |8300 |8287 |
> 0 | 1 | 1 | 0 | 1.460359e-02 | 0.000000e+00 | Inf | unknown
> 0.9s| 1 | 0 | 1328 | - | 50M | 0 |1580 |8297 |8287 |
> 0 | 2 | 1 | 0 | 1.460359e-02 | 0.000000e+00 | Inf | unknown
> 0.9s| 1 | 0 | 1328 | - | 50M | 0 |1580 |8297 |8287 |
> 0 | 2 | 1 | 0 | 1.460359e-02 | 0.000000e+00 | Inf | unknown
> 0.9s| 1 | 0 | 1328 | - | 50M | 0 |1580 |8277 |8287 |
> 0 | 4 | 1 | 0 | 1.460359e-02 | 0.000000e+00 | Inf | unknown
> 1.0s| 1 | 0 | 1333 | - | 50M | 0 |1580 |8277 |8289 |
> 2 | 5 | 1 | 0 | 1.319712e-02 | 0.000000e+00 | Inf | unknown
> 1.0s| 1 | 0 | 1368 | - | 51M | 0 |1580 |8277 |8293 |
> 6 | 6 | 1 | 0 | 1.158039e-02 | 0.000000e+00 | Inf | unknown
> 1.5s| 1 | 0 | 1864 | - | 52M | 0 |1580 |8278 |8293 |
> 6 etc...
>
> Since the column node shows 0 and 1, does this mean SCIP spent 1.5
> seconds on solving 2 LPs ?
>
> Is there any detailed docs about what each column mean, in particular
> the "sepa", "confs" and "strbr" (the other ones I have no doubts).
>
> Thanks!
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip
>
More information about the Scip
mailing list