[Ipopt] re. local minima

Stefan Vigerske stefan at math.hu-berlin.de
Mon Feb 15 01:27:27 EST 2016


Ipopt should stop when having proven local optimality of a solution. 
Maybe check that your derivatives are implemented correctly (e.g., by 
using Ipopt's derivative checker) and your problem is well scaled and 
functions are smooth.
Regarding settings, you can try whether increasing tolerances, 
especially for dual infeasibility and complementarity, would help.


On 02/14/2016 01:07 PM, Nithin Kurup wrote:
> Hi,
> I am relatively new to optimization. I am optimizing 2 variables for a
> trial and noticed that in my problem, occasionally the objective function
> falls into a local minima, (even though there were sort of global minimum
>   achieved in the previous iterations) and stays in this for a very long
> time say, around 1500 iterations ( believe its trying to find a new search
> direction and step size). Is there any settings I can use to avoid this and
> save computational time.
> Looking forward to your kind help.
> Nithin
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt


More information about the Ipopt mailing list