[Ipopt] Problems passing derivatives information to IPOPT

Miguel Angel Salazar de Troya salazardetroya at gmail.com
Thu Jan 8 08:50:41 EST 2015


In the derivatives test provided by Ipopt, my inequalities derivatives
failed. The derivatives that I provided are wrong, but they are wrong when
they are displayed in the derivative test. Inside of eval_jac_g() they
match with the finite difference, therefore, the problem must be in the way
these are passed to IpOpt.

I have 90 variables and just one constraint whose derivative is a constant
for all the variables. The iRow and jCol structures are therefore as
iRow[0] = iRow[1] = ... iRow[89] = 0

jCol[0] = 0;
jCol[1] = 1;
jCol[89] = 89;

I fill the array values with the corresponding constant "C"
value[0] = value[1] = ... value[89] =  C

I print this array to verify everything is correct. These values are not
the same than in the derivative test. Interestingly, only the first
variable of the derivative has a non-zero value and it's the only one with
the ``v''  next to it. I must be forgetting something pretty basic, but I'm
not able to find it. I would appreciate any type of help. Please let me
know if I should provide more details.

Just to double check: In the derivatives test, the first column corresponds
to the user derivatives and the second column to the finite difference
test, right?

Thanks in advance
*Miguel Angel Salazar de Troya*
Graduate Research Assistant
Department of Mechanical Science and Engineering
University of Illinois at Urbana-Champaign
(217) 550-2360
salaza11 at illinois.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/ipopt/attachments/20150108/81a6636f/attachment.html>

More information about the Ipopt mailing list