[Scip] SCIP interface with AMPL
Cristiano Arbex
crisarbex at gmail.com
Mon Sep 30 16:55:16 CEST 2013
Hi again,
I've been using the NEOS server to test SCIP and I've been impressed with
its speed. I tried solving 36 nonlinear instances and the largest one can
be downloaded here:
https://docs.google.com/file/d/0B4Rt5z2PACNndUZUc184dVBxQVU/edit?usp=sharing
It takes NEOS SCIP 6 seconds to solve it. I am using the same scip.set
parameters file that can be found on NEOS. I installed the SCIP
optimization suite following the steps in the link provided in the first
message of this thread.
I know my PC is slower than the NEOS server but SCIP on my PC has been
solving the same instance for 1200s. In the output file I get this message:
--------------
Presolving Time: 4.20
Quadratic constraint handler does not have LAPACK for eigenvalue
computation. Will assume that matrices (with size > 2x2) are indefinite.
time | node | left |LP iter|LP it/n| mem |mdpt |frac |vars |cons |cols
|rows |cuts |confs|strbr| dualbound | primalbound | gap
6.5s| 1 | 0 | 1111 | - | 21M| 0 | 6 |4268 |3217 |4268
|3196 | 0 | 15 | 0 | 0.000000e+00 | -- | Inf
6.7s| 1 | 0 | 1563 | - | 21M| 0 | 6 |4268 |3217 |4268
|3196 | 0 | 15 | 0 | 0.000000e+00 | -- | Inf
6.8s| 1 | 0 | 1575 | - | 21M| 0 | 6 |4268 |3217 |4268
|3207 | 11 | 15 | 0 | 0.000000e+00 | -- | Inf
--------------
While NEOS output file looks like this:
--------------
Presolving Time: 2.54
time | node | left |LP iter|LP it/n| mem |mdpt |frac |vars |cons
|cols |rows |cuts |confs|strbr| dualbound | primalbound | gap
3.8s| 1 | 0 | 1060 | - | 21M| 0 | 6 |4268 |3217
|4268 |3196 | 0 | 15 | 0 | 0.000000e+00 | -- | Inf
******************************************************************************
This program contains Ipopt, a library for large-scale nonlinear optimization.
Ipopt is released as open source code under the Eclipse Public License (EPL).
For more information visit http://projects.coin-or.org/Ipopt
******************************************************************************
q 6.2s| 1 | 0 | 1138 | - | 21M| 0 | - |4268 |3217
|4268 |3196 | 0 | 15 | 0 | 0.000000e+00 | 0.000000e+00 | 0.00%
6.2s| 1 | 0 | 1138 | - | 21M| 0 | - |4268 |3217
|4268 |3196 | 0 | 15 | 0 | 0.000000e+00 | 0.000000e+00 | 0.00%
--------------
Do you know why I cannot reproduce such results? Might I be missing
something, like the IPOPT nonlinear solver?
Thank you very much,
Cristiano
On Sat, Sep 28, 2013 at 4:30 PM, Cristiano Arbex <crisarbex at gmail.com>wrote:
> Thank you very much, Stefan, that was very helpful.
>
>
> On Sat, Sep 28, 2013 at 2:06 PM, Stefan Vigerske <stefan at math.hu-berlin.de
> > wrote:
>
>> Hi,
>>
>> what is not supported so far is to set SCIP options via the AMPL script
>> language.
>> However, AMPL/SCIP reads by default SCIP options from the file scip.set,
>> if present.
>>
>> Alternatively, one can specify a SCIP settings file as argument to the
>> AMPL/SCIP binary. However, that would require to write out the AMPL model
>> as .nl file first and to run the binary explicitly, e.g.,
>> ./bin/scipampl foo.nl myoptions.set
>>
>> Btw, adding "-i" to the command line arguments gives you the SCIP
>> interactive shell.
>>
>> Hope that helps,
>> Stefan
>>
>>
>>
>> On 09/28/2013 01:55 PM, Cristiano Arbex wrote:
>>
>>> I am trying to use SCIP with its interface to AMPL, which I installed on
>>> Linux via this following tutorial:
>>>
>>> http://zverovich.net/2012/08/**07/using-scip-with-ampl.html<http://zverovich.net/2012/08/07/using-scip-with-ampl.html>
>>>
>>>
>>> In the tutorial it is said that no SCIP options are currently supported.
>>> Do
>>> you know if there is a way to use options with AMPL or if there is an
>>> alternate way to do this, for example, by presenting an options file?
>>>
>>> Thank you very much,
>>> Cristiano
>>>
>>>
>>>
>>> ______________________________**_________________
>>> Scip mailing list
>>> Scip at zib.de
>>> http://listserv.zib.de/**mailman/listinfo/scip<http://listserv.zib.de/mailman/listinfo/scip>
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20130930/d0ef9a79/attachment.html>
More information about the Scip
mailing list