-nan(ind) #564

TheSimplyRed · 2022-03-21T19:12:18Z

TheSimplyRed
Mar 21, 2022

Hi,
I got something strange. During some iterations I have x values that becomes -nan(ind)!
What can be wrong in such cases?

Answered by svigerske

Some computations lead to nan.

First try turning on option check_derivatives_for_naninf and make sure you use Ipopt 3.14.

If that doesn't help, compile Ipopt with --enable-debug --with-ipopt-checklevel=2 (https://coin-or.github.io/Ipopt/INSTALL.html#CONFIGURE_FLAGS) and/or set a high print_level to maybe see which operations cause the nan's.

svigerske · 2022-03-22T08:02:25Z

Some computations lead to nan.

First try turning on option check_derivatives_for_naninf and make sure you use Ipopt 3.14.

If that doesn't help, compile Ipopt with --enable-debug --with-ipopt-checklevel=2 (https://coin-or.github.io/Ipopt/INSTALL.html#CONFIGURE_FLAGS) and/or set a high print_level to maybe see which operations cause the nan's.

0 replies