[Bonmin] [*****SPAM*****] Re: error in running bonmin

[Giuseppe Aprea]

Thanks for your suggestions. I am solving a
mixed integer nonlinear(quadratic) problem.

tot number of variables: 3911
number of binary variables: 560
linear inequality constraints: 1125
linear equality constraints: 2227
nonlinear equality constraint: 1


Do you think Bonmin can handle it? I am using Mumps at the moment because
people from HSL have not answered me yet.

cheers,

giuseppe



On Wed, Jul 1, 2009 at 2:06 AM, Edson Valle<edsoncv at gmail.com> wrote:
>   Yes, just a little correction derivative_test second-order (see
> underscore, see http://www.coin-or.org/Ipopt/documentation/node164.html, it
> is actually a Ipopt option).  This option will print only derivatives which
> error is greater than the tolerance ( see
> http://www.coin-or.org/Ipopt/documentation/node166.html).
>
> You can check also first derivatives and also the options:
>
> derivative_test_perturbation: perturbation size in the finite diff.
> derivative checker.
>
> derivative_test_print_all : print all derivatives even if no error is found.
>
> Generally a valgrind associated with these tests can find the problem, thats
> my experience :) .
>
> By the way, you are using bonmin to solve what kind of problem?
>
> cheers
>
> 2009/6/30 Giuseppe Aprea <giuseppe.aprea at gmail.com>
>>
>> Thank a lot for your suggestions; by the way, is there a way to check
>> first and second derivatives?
>>
>> maybe putting
>>
>> "derivative test second-order"
>>
>> in Mybonmin.opt?
>>
>> cheers
>>
>> giuseppe
>>
>> On Tue, Jun 30, 2009 at 10:31 PM, Edson Cordeiro do
>> Valle<edsoncv at gmail.com> wrote:
>> > Hello    I usually use valgrind to detect leaks and others issues in
>> > Ipopt
>> > and bonmin with the folowing options:
>> > valgrind --tool=memcheck --leak-check=yes --log-file=val
>> >  yourexecfilename
>> > The results will be written to the file name "val".
>> > Make sure that you build the ipopt and bonmin with debug flag, this will
>> > make the line number of the problem available. You your application that
>> > uses Ipopt/bonmin has debug symbols, you can have better outputs in
>> > valgrind.
>> > I hope this helps, good luck.
>> > Regards
>> > Giuseppe Aprea escreveu:
>> >>
>> >> I see.
>> >>
>> >> Any suggestion for a step by step debugging? I guess I am going out of
>> >> bound somewhere
>> >> but it's difficult for me to check; I usually write programs in
>> >> python.....
>> >>
>> >> maybe some compiler options? valgrind (in case, which options)?
>> >>
>> >> thanks
>> >>
>> >> giuseppe
>> >>
>> >> On Mon, Jun 29, 2009 at 10:31 AM, Stefan Vigerske<vigerske at zib.de>
>> >> wrote:
>> >>
>> >>>
>> >>> Hi,
>> >>>
>> >>> the problem seem to be in your code, so it's hard to tell what is
>> >>> going
>> >>> wrong without seeing it.
>> >>> new_x is a flag that tells you whether the evaluation methods have
>> >>> already been called for the current x. It should be explained in the
>> >>> Ipopt documentation of the TNLP.
>> >>>
>> >>> Stefan
>> >>>
>> >>> Giuseppe Aprea schrieb:
>> >>>
>> >>>>
>> >>>> Hi,
>> >>>>
>> >>>> I am using Bonmin (retrieved by svn last week) with Ipopt+ MUMPS
>> >>>> (unfortunately I have not received any answer from MA27 developers)
>> >>>> I get this error while running my MINLP.
>> >>>>
>> >>>> It seems it Cannot access memory for new_x. My problem is that I
>> >>>> could not find anything about new_x in the documentation. I am just
>> >>>> using
>> >>>> the same eval_jac_g function definition of the cpp example directory
>> >>>> of
>> >>>> bonmin
>> >>>> which works fine. Of course the values of the jacobian are different.
>> >>>>
>> >>>> Does anybody have an idea about the possible error?
>> >>>>
>> >>>> Core was generated by `./CppExample'.
>> >>>> Program terminated with signal 11, Segmentation fault.
>> >>>> [New process 4384]
>> >>>> #0  0x0804e96d in MyTMINLP::eval_jac_g (this=0x805ca38, n=3911,
>> >>>> x=0x0,
>> >>>> new_x=Cannot access memory at address 0xb9652394
>> >>>> ) at MyTMINLP.cpp:408
>> >>>> 408                          Number* values)
>> >>>> (gdb) bt
>> >>>> #0  0x0804e96d in MyTMINLP::eval_jac_g (this=0x805ca38, n=3911,
>> >>>> x=0x0,
>> >>>> new_x=Cannot access memory at address 0xb9652394
>> >>>> ) at MyTMINLP.cpp:408
>> >>>> #1  0xb7ea51b6 in Bonmin::TMINLP2TNLP::eval_jac_g () from
>> >>>> /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #2  0xb7a9f461 in Ipopt::TNLPAdapter::GetSpaces () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #3  0xb7cc250f in Ipopt::OrigIpoptNLP::InitializeStructures () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #4  0xb7c86b9a in Ipopt::IpoptData::InitializeDataStructures () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #5  0xb7c2c569 in
>> >>>> Ipopt::DefaultIterateInitializer::SetInitialIterates
>> >>>> () from /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #6  0xb7c4e08f in Ipopt::IpoptAlgorithm::InitializeIterates () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #7  0xb7c4e7e7 in Ipopt::IpoptAlgorithm::Optimize () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #8  0xb7a7b718 in Ipopt::IpoptApplication::call_optimize () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #9  0xb7a8094e in Ipopt::IpoptApplication::OptimizeNLP () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #10 0xb7a80e9a in Ipopt::IpoptApplication::OptimizeNLP () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #11 0xb7a80fdd in Ipopt::IpoptApplication::OptimizeTNLP () from
>> >>>> /home/gaprea/usr/local/lib/libipopt.so.0
>> >>>> #12 0xb7ebcd85 in Bonmin::IpoptSolver::OptimizeTNLP () from
>> >>>> /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #13 0xb7e9bc4d in Bonmin::OsiTMINLPInterface::solveAndCheckErrors ()
>> >>>> from /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #14 0xb7e9da18 in Bonmin::OsiTMINLPInterface::initialSolve () from
>> >>>> /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #15 0xb79c861e in CbcModel::initialSolve () from
>> >>>> /home/gaprea/usr/local/lib/libCbc.so.0
>> >>>> #16 0xb7e21b37 in Bonmin::Bab::branchAndBound () from
>> >>>> /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #17 0xb7e20e17 in Bonmin::Bab::operator() () from
>> >>>> /home/gaprea/usr/local/lib/libbonmin.so.0
>> >>>> #18 0x0804b709 in main () at MyBonmin.cpp:79
>> >>>>
>> >>>> thanks in advance,
>> >>>>
>> >>>> giuseppe
>> >>>>
>> >>>> _______________________________________________
>> >>>> Bonmin mailing list
>> >>>> Bonmin at list.coin-or.org
>> >>>> http://list.coin-or.org/mailman/listinfo/bonmin
>> >>>>
>> >>>>
>> >>>
>> >>>
>> >>
>> >>
>> >> _______________________________________________
>> >> Bonmin mailing list
>> >> Bonmin at list.coin-or.org
>> >> http://list.coin-or.org/mailman/listinfo/bonmin
>> >>
>> >
>> >
>
>
>
> --
>       Edson Valle
> edsoncv at gmail.com
>