[Bonmin] a problem of bonmin
Pierre Bonami
pierre.bonami at gmail.com
Fri Mar 16 18:05:16 EDT 2007
Dear Roy,
Sorry there was a typo in the bonmin.opt.
The value for bonmin.cutoff should have been 1e100 and not -1e100
(this value makes the problem infeasible).
I upload a corrected bonmin.opt in the stable branch and attach the
correction.
Best,
Pierre
On Mar 16, 2007, at 2:02 PM, Roy Chen wrote:
>
> Dear all
>
> I found another problem regarding Bonmin. When I tried the example
> as attached without putting the default bonmin.opt file in the
> folder, I got the right solution as following:
> Finished
> /tmp/at17406
> bonmin: Optimal
> : v b tL :=
> 1 8.00637 5.9395 1.16315
> 2 7.54515 6.6468 1.22378
> 3 7.58816 6.58964 1.82455
> 4 7.87057 6.45881 1.22378
> 5 7.75279 6.26422 1.30833
> 6 7.65435 . .
> ;
>
> y [*,*] (tr)
> : 1 2 3 4 :=
> 1 0 1 0 0
> 2 0 1 0 0
> 3 0 0 1 0
> 4 0 1 0 0
> 5 1 0 0 0
> 6 1 0 0 0
> ;
>
> cost = 285506
>
>
> But when I put the bonmin.opt in defaut setting in the same folder
> where I put the .mod and .x file, I got the different output, which
> is obvious wrong and shares the similar problem as I mentioned in
> my previous email, such as:
> Finished
> /tmp/at17442
> bonmin: Infeasible problem
> : v b tL :=
> 1 0 0 0
> 2 0 0 0
> 3 0 0 0
> 4 0 0 0
> 5 0 0 0
> 6 0 . .
> ;
>
> y [*,*] (tr)
> : 1 2 3 4 :=
> 1 0 0 0 0
> 2 0 0 0 0
> 3 0 0 0 0
> 4 0 0 0 0
> 5 0 0 0 0
> 6 0 0 0 0
> ;
>
> cost = 1500
>
> Please also see the attachment as the .mod, .x and bonmin.opt files
> I tried.
>
> Best
>
> Roy
>
>
>
>
> On 3/14/07, Pierre Bonami <pierre.bonami at gmail.com > wrote:
> Dear Roy,
> I received your email and I have been looking into this but with no
> success yet.
> Best,
> Pierre
>
> On Mar 14, 2007, at 11:47 AM, Roy Chen wrote:
>
>> Dear Pierre
>>
>> I am just wondering if you have found out the reason with my
>> Bonmin not being able to return the feasible solution?
>>
>> Best
>>
>> Roy
>>
>>
>>
>>
>> On 3/13/07, Pierre Bonami <pierre.bonami at gmail.com > wrote:
>> Dear Roy,
>> Could you rerun the exact same problem with print_level 4 and
>> bonmin.bb_log_level 3?
>> Cheers,
>> Pierre
>>
>> On Mar 13, 2007, at 12:27 PM, Roy Chen wrote:
>>
>>> Dear Pierre
>>> I changed one of those two variables, y into integer variable,
>>> but it seems still not working well.
>>>
>>> bonmin:
>>> ********************************************************************
>>> **********
>>> This program contains Ipopt, a library for large-scale nonlinear
>>> optimization.
>>> Ipopt is released as open source code under the Common Public
>>> License (CPL).
>>> For more information visit http://projects.coin-or.org/
>>> Ipopt
>>> ********************************************************************
>>> **********
>>>
>>> Restoration phase is called at point that is almost feasible,
>>> with constraint violation 3.552714e-15. Abort.
>>> Restoration phase is called at point that is almost feasible,
>>> with constraint violation 3.552714e-15. Abort.
>>> Restoration phase is called at point that is almost feasible,
>>> with constraint violation 5.062617e-14. Abort.
>>> Restoration phase is called at point that is almost feasible,
>>> with constraint violation 4.440892e-15. Abort.
>>> OA0001I Solved NLP in 6 iterations, found a feasible solution of
>>> value 7.417879.
>>> Cbc0013I At root node, 0 cuts changed objective from 7.41788 to
>>> 7.41788 in 1 passes
>>> Cbc0014I Cut generator 0 (Outer Approximation Supporting
>>> Hyperplanes for NLP optimum) - 0 row cuts (0 active), 0 column cuts
>>> Cbc0014I Cut generator 1 (GMI) - 0 row cuts (0 active), 0 column
>>> cuts
>>> Cbc0014I Cut generator 2 (covers) - 0 row cuts (0 active), 0
>>> column cuts
>>> Cbc0014I Cut generator 3 (MIR) - 0 row cuts (0 active), 0 column
>>> cuts
>>> Cbc0014I Cut generator 4 (Outer Approximation local enumerator) -
>>> 1 row cuts (0 active), 0 column cuts
>>> Cbc0014I Cut generator 5 (Outer Approximation feasibility
>>> checker) - 0 row cuts (0 active), 0 column cuts
>>> Cbc0001I Search completed - best objective 1e+50, took 0
>>> iterations and 0 nodes ( 0.01 seconds)
>>> Finished
>>> /tmp/at7669
>>> bonmin: Infeasible problem
>>> x = 0
>>>
>>> y = 0
>>>
>>> Z = 0
>>>
>>>
>>> Best
>>> Roy
>>>
>>> On 3/13/07, Pierre Bonami <pierre.bonami at gmail.com > wrote:
>>> Dear Roy,
>>> The problem (as seen from the output you attached from the
>>> ticket). That Bonmin did not handle properly return codes when no
>>> integer variable was present in the model.
>>> I fixed that in the trunk version. I can try to fix it in the
>>> stable version and make a new release.
>>>
>>>
>>> In any case it should work better if you put integer variables in
>>> the model (and otherwise it can only be worst than Ipopt).
>>>
>>>
>>> Best regards,
>>> Pierre
>>>
>>> On Mar 13, 2007, at 11:26 AM, Roy Chen wrote:
>>>
>>>> Dear Pierre
>>>> I downloaded Bonmin from http://www.coin-or.org/download/source/
>>>> Bonmin/ in last December, I believe, therefore, it should be
>>>> stable version. As a matter of fact, I don't quite know how to
>>>> download truck version.
>>>> Best
>>>> Roy
>>>>
>>>>
>>>> On 3/13/07, Pierre Bonami <pierre.bonami at gmail.com > wrote:
>>>> Dear Roy,
>>>> Are you using stable or trunk?
>>>> I assumed from previous messages (64 bits and Ampl) that you
>>>> were using trunk.
>>>> Best,
>>>> Pierre
>>>> On Mar 12, 2007, at 12:41 AM, Roy Chen wrote:
>>>>
>>>>> Hello everyone
>>>>>
>>>>> I tried the following problem:
>>>>>
>>>>> function Integral1(Reals,Reals,Reals);
>>>>> var x ;
>>>>> var y ;
>>>>> minimize Z:tan(x)+y*Integral1(x,0,1);
>>>>> subject to constrain:x+y>=5;
>>>>> subject to constraint:2*x+3*y<=27;
>>>>> option solver ipopt;
>>>>> solve;
>>>>> display x;
>>>>> display y;
>>>>> display Z;
>>>>>
>>>>> When the solver option was chosen "ipopt" as above, I got the
>>>>> following solution:
>>>>> EXIT: Optimal Solution Found.
>>>>> Ipopt 3.2.2: Optimal Solution Found
>>>>> x = -12
>>>>> y = 17
>>>>> Z = 7.41788
>>>>> When the solver option was chosen "bonmin" and every parameter
>>>>> in bonmin remains default, I got the following:
>>>>> bonmin:
>>>>> ******************************************************************
>>>>> ************
>>>>> This program contains Ipopt, a library for large-scale
>>>>> nonlinear optimization.
>>>>> Ipopt is released as open source code under the Common Public
>>>>> License (CPL).
>>>>> For more information visit http://projects.coin-or.org/
>>>>> Ipopt
>>>>> ******************************************************************
>>>>> ************
>>>>>
>>>>> Ipopt exited with error code -2 Restoration Failed
>>>>> Error in opening Names file (probably does not exists)can't
>>>>> open /tmp/at2970.sol
>>>>>
>>>>> Note: Integral1() is the external function I defined by myself.
>>>>> Actually I tried all of the four algorithms in bonmin, one of
>>>>> them works.
>>>>>
>>>>> Why the two outputs are so different? Anyone has some idea?
>>>>>
>>>>> Thanks
>>>>>
>>>>> Roy
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Bonmin mailing list
>>>>> Bonmin at list.coin-or.org
>>>>> http://list.coin-or.org/mailman/listinfo/bonmin
>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>
>
> <batch.mod>
> <batch.x>
> <bonmin.opt>
> _______________________________________________
> Bonmin mailing list
> Bonmin at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/bonmin
-------------- next part --------------
Skipped content of type multipart/mixed
More information about the Bonmin
mailing list