[SCIP] TR: scip crash
Jakob Witzig
witzig at zib.de
Thu Jul 21 16:03:56 CEST 2016
Hi Nicolas,
I tried one of the commercial MIP solvers and checked the provided
"optimal" solution for feasibility within SCIP. The solution was
rejected due to constraint violations. To achieve solutions you can work
with you should definitively change your model.
Just as a remark, you should always be a bit suspicious whenever a
solver has highly numerical troubles and another gives you an optimal
solution very fast.
Cheers,
Jakob
On 21.07.2016 14:01, nicolas.derhy at engie.com wrote:
> Ok, thank you for your answer.
> I have similar problem with Coin (CBC) with this kind of data.
> I will try to modify my model slightly.
>
> -----Message d'origine-----
> De : Ambros Gleixner [mailto:gleixner at zib.de]
> Envoyé : jeudi 21 juillet 2016 13:07
> À : zjuzw at sohu.com; Derhy Nicolas (ENGIE) <nicolas.derhy at engie.com>; scip <scip at zib.de>
> Objet : Re: [SCIP] TR: scip crash
>
> Hi Nicolas,
>
> The problem is indeed that SoPlex fails to solve one LP because of the huge coefficient range in the matrix.
>
> There seems to be no easy workaround just changing parameters. We will look into it, but you should not expect a quick fix.
>
> Ideally you should try to improve the numerical properties of your model.
>
> Best,
> Ambros
>
>
>
> Am 05.07.2016 um 11:02 schrieb zjuzw at sohu.com:
>> Hello nicolas:
>>
>> Fortunately,it is a mixed 0-1 integer programming problem,and
>> successfully solved by CPLEX.
>>
>> The result is as follows:
>>
>> Welcome to IBM(R) ILOG(R) CPLEX(R) Interactive Optimizer 12.6.3.0
>> with Simplex, Mixed Integer & Barrier Optimizers
>> 5725-A06 5725-A29 5724-Y48 5724-Y49 5724-Y54 5724-Y55 5655-Y21
>> Copyright IBM Corp. 1988, 2015. All Rights Reserved.
>>
>> Type 'help' for a list of available commands.
>> Type 'help' followed by a command name for more information on
>> commands.
>>
>> CPLEX> read bug.mps
>> Selected objective sense: MINIMIZE
>> Selected objective name: C
>> Selected RHS name: RHS1
>> Selected range name: RNG1
>> Selected bound name: BND1
>> Problem 'bug.mps' read.
>> Read time = 0.16 sec. (0.79 ticks)
>> CPLEX> optimize
>> Tried aggregator 5 times.
>> MIP Presolve eliminated 1629 rows and 1354 columns.
>> MIP Presolve modified 1110 coefficients.
>> Aggregator did 268 substitutions.
>> Reduced MIP has 307 rows, 276 columns, and 919 nonzeros.
>> Reduced MIP has 119 binaries, 0 generals, 0 SOSs, and 0 indicators.
>> Presolve time = 0.06 sec. (4.38 ticks) Found incumbent of value
>> 1.0991956e+018 after 0.17 sec. (4.97 ticks) Probing fixed 13 vars,
>> tightened 21 bounds.
>> Probing time = 0.00 sec. (0.09 ticks)
>> Tried aggregator 2 times.
>> MIP Presolve eliminated 103 rows and 101 columns.
>> MIP Presolve modified 79 coefficients.
>> Aggregator did 6 substitutions.
>> Reduced MIP has 197 rows, 169 columns, and 578 nonzeros.
>> Reduced MIP has 67 binaries, 0 generals, 0 SOSs, and 0 indicators.
>> Presolve time = 0.00 sec. (0.60 ticks) Probing fixed 0 vars, tightened
>> 1 bounds.
>> Probing time = 0.00 sec. (0.05 ticks)
>> Tried aggregator 1 time.
>> MIP Presolve eliminated 10 rows and 12 columns.
>> MIP Presolve modified 13 coefficients.
>> Reduced MIP has 187 rows, 157 columns, and 554 nonzeros.
>> Reduced MIP has 63 binaries, 0 generals, 0 SOSs, and 0 indicators.
>> Presolve time = 0.00 sec. (0.32 ticks) Probing time = 0.00 sec. (0.04
>> ticks) Clique table members: 14.
>> MIP emphasis: balance optimality and feasibility.
>> MIP search method: dynamic search.
>> Parallel mode: deterministic, using up to 4 threads.
>> Root relaxation solution time = 0.00 sec. (0.39 ticks)
>>
>> Nodes Cuts/
>> Node Left Objective IInf Best Integer Best Bound
>> ItCnt Gap
>>
>> * 0+ 0 1.09920e+018 8.18653e+017
>> 25.52%
>> * 0+ 0 8.50858e+017
>> 8.18653e+017 3.78%
>> 0 0 8.22495e+017 4 8.50858e+017 8.22495e+017
>> 8 3.33%
>> * 0+ 0 8.25697e+017
>> 8.22495e+017 0.39%
>> 0 0 8.22495e+017 5 8.25697e+017 Cuts: 16
>> 20 0.39%
>> * 0+ 0 8.22495e+017
>> 8.22495e+017 0.00%
>> 0 0 cutoff 8.22495e+017 8.22495e+017
>> 20 0.00%
>> Elapsed time = 0.23 sec. (11.98 ticks, tree = 0.00 MB, solutions = 3)
>>
>> Clique cuts applied: 3
>> Implied bound cuts applied: 5
>> Mixed integer rounding cuts applied: 2 Gomory fractional cuts
>> applied: 4
>>
>> Root node processing (before b&c):
>> Real time = 0.25 sec. (12.05 ticks)
>> Parallel b&c, 4 threads:
>> Real time = 0.00 sec. (0.00 ticks)
>> Sync time (average) = 0.00 sec.
>> Wait time (average) = 0.00 sec.
>> ------------
>> Total (root+branch&cut) = 0.25 sec. (12.05 ticks)
>>
>> Solution pool: 4 solutions saved.
>>
>> MIP - Integer optimal solution: Objective = 8.2249485120e+017
>> Solution time = 0.61 sec. Iterations = 20 Nodes = 0
>> Deterministic time = 12.06 ticks (19.83 ticks/sec)
>>
>> But the global optimization solvers,such as SCIP,BARON,COUENNE,can't
>> bring the same result as CPLEX.
>>
>> Does it mean that Soplex is inferior to CPLEX,which serves as the
>> default linear programming solver in SCIP?
>>
>>
>>
>>
>>
>> Best Wishes,
>>
>> From Wei.
>>
>>
>>
>>
>> ----------------------------------------------------------------------
>> --
>>
>> <http://score.mail.sohu.com/?ref=mail_tailad>
>>
>>
>>
>>
>>
>> _______________________________________________
>> Scip mailing list
>> Scip at zib.de
>> http://listserv.zib.de/mailman/listinfo/scip
>>
> --
> Ambros Gleixner, Zuse Institute Berlin, http://www.zib.de/gleixner
> ENGIE Mail Disclaimer: http://www.engie.com/disclaimer/disclaimer-fr.html
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip
--
Jakob Witzig
Zuse Institute Berlin (ZIB)
Division Mathematical Optimization and Scientific Information
Research Group Mathematical Optimization Methods
Takustrasse 7
14195 Berlin
Tel. : +49 (0)30 84185-416
Fax : +49 (0)30 84185-269
email: witzig at zib.de
More information about the Scip
mailing list