[Dip] "Not a mixed-integer problem" message when using DIP with CPLEX

Matthew Galati matthew.galati at gmail.com
Mon Feb 21 10:25:42 EST 2011


This is fixed at trunk. I just needed to switch the problem type to MIP upon
construction of subproblem OSI. It should have had no effect on results,
since the solve call automatically does the switch again (if needed).

Thanks,
Matt



On Wed, Jan 26, 2011 at 3:20 PM, Matthew Galati <matthew.galati at gmail.com>wrote:

> I've seen this for a while and well after the fixes mentioned. It has no
> real effect on DIP as far as I can tell. But, I'll investigate when I have
> time.
>
>
>
> On Wed, Jan 26, 2011 at 3:18 PM, Ted Ralphs <ted at lehigh.edu> wrote:
>
>> We've seen messages like this with CPLEX for a long time in many
>> contexts (not just DIP). Unlike most other solvers, CPLEX makes a very
>> strong distinction between an LP and a MIP. Although conceptually, the
>> only difference is whether there exist variables that are marked as
>> integer type or not, CPLEX takes the "problem type," as set by the
>> user, as strongly dictating what can and cannot be done when
>> manipulating the model. CPLEX also erases all solution information
>> when switching problem type. The use of Cgl actually requires the
>> problem type to be switched back and forth and this causes problems. I
>> thought we had actually fixed all of these problems over time, so if
>> you are seeing this error now, it should be an indication of a true
>> problem.
>>
>> Ted
>>
>> On Wed, Jan 26, 2011 at 1:49 PM, Matthew Galati
>> <matthew.galati at gmail.com> wrote:
>> > I know about it.... it doesn't seem to effect anything - except that it
>> is
>> > annoying.
>> >
>> > It is on my ToDo list to investigate why that message appears. I think
>> it
>> > has something to do with the odd design of OsiCpx.
>> >
>> >
>> >
>> > On Wed, Jan 26, 2011 at 1:46 PM, kim <kim at hamilton-vejlin.dk> wrote:
>> >>
>> >> I just set up DIP to use CPLEX instead of CBC and it works well but I
>> am
>> >> getting some odd messages at the beginning of the program run.
>> >>
>> >> As one of the first things I get a string of messages all saying:
>> >> "CPX0000  CPLEX Error  3003: Not a mixed-integer problem."
>> >>
>> >> This happens before the search is actually started and there is always
>> >> one "error 3003" for each block. It doesn't seem to have an effect at
>> >> all and the program behaves as expected, but it seems odd.
>> >>
>> >> Does anyone have in idea what those messages are about and if I should
>> >> see it as a sign I have not set up the DIP/CPLEX combination correctly?
>> >>
>> >> -Kim
>> >>
>> >> _______________________________________________
>> >> Dip mailing list
>> >> Dip at list.coin-or.org
>> >> http://list.coin-or.org/mailman/listinfo/dip
>> >
>> >
>> > _______________________________________________
>> > Dip mailing list
>> > Dip at list.coin-or.org
>> > http://list.coin-or.org/mailman/listinfo/dip
>> >
>>
>>
>>
>> --
>> Dr. Ted Ralphs
>> Associate Professor, Lehigh University
>> (610) 628-1280
>> ted 'at' lehigh 'dot' edu
>> coral.ie.lehigh.edu/~ted <http://coral.ie.lehigh.edu/%7Eted>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.coin-or.org/pipermail/dip/attachments/20110221/a3c0747c/attachment.html 


More information about the Dip mailing list