[Ipopt] contraint violation - e+14 !?!

[Andreas Waechter]

Hi Denis,

Ipopt can also (since recently) be asked to start checking the derivatives 
starting at a particular index (see selected output of
"ipopt --print-options" below).  Using this, you can check a few 
derivatives at random, as you suggested.

Regards,

Andreas

---------

derivative_test               ("none")
    Enable derivative checker
      If this option is enabled, a (slow!) derivative test will be performed
      before the optimization.  The test is performed at the user provided
      starting point and marks derivative values that seem suspicious
    Possible values:
     - none                    [do not perform derivative test]
     - first-order             [perform test of first derivatives at starting
                                point]
     - second-order            [perform test of first and second derivatives at
                                starting point]
     - only-second-order       [perform test of second derivatives at starting
                                point]

derivative_test_first_index           -2 <= (         -2) <  +inf
    Index of first quantity to be checked by derivative checker
      If this is set to -2, then all derivatives are checked.  Otherwise, for
      the first derivative test it specifies the first variable for which the
      test is done (counting starts at 0).  For second derivatives, it
      specifies the first constraint for which the test is done; counting of
      constraint indices starts at 0, and -1 refers to the objective function
      Hessian.


On Wed, 11 Nov 2009, Denis Davydov wrote:

> Dear Sebastian,
>
> Yes, I provide exact Hessian.
>
> Thanks for the idea of checking just few dimensions at random, i was
> thinking to check Hessian with build-in capabilities of Ipopt but it just
> takes too much time.
>
> Best regards,
> Denis.
>
> <quote who="Sebastian Nowozin">
>> Dear Denis Davydov,
>>
>> just from my experience with Ipopt: seeing 500 iterations is unusual.
>> I have only seen such behaviour when the limited-memory quasi-Newton
>> approximation to the Hessian matrix is used, but not with the exact
>> Hessian.  Do you provide the exact Hessian?
>>
>> Also, this might be obvious, but did you ensure your gradient and
>> Hessian computations are correct by checking them against
>> finite-differences?  (When computing the objective takes too long for
>> this to be feasible you can just test a few dimensions sampled at
>> random.)
>>
>> Regards,
>> Sebastian
>>
>
>
> -- 
> Denis Davydov
> Department of Mechanics
> Faculty of Civil Engineering
> Czech Technical University
> Thakurova 7, 166 29 Prague 6, Czech Republic
> phone: +420-728478669
> email: denis.davydov at fsv.cvut.cz
>
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>
>