[Ipopt] CPU time limit setting IPOPT

Stefan Vigerske stefan at math.hu-berlin.de
Tue Dec 4 08:25:45 EST 2012


Ipopt does not guarantee that it will not exceed the time limit by a 
single millisecond.
The time limit is only checked at certain plances (maybe just once per 
iteration) and Ipopt stops when it finds the limit exceeded. By how much 
it is exceeded depends on the effort that Ipopt has to spend per iteration.

That should be the explanation, but I don't have an easy solution. I 
think Ipopt is just not made for real-time applications.


On 12/04/2012 02:01 PM, Martijn Disse wrote:
> Hi Everyone,
> I am using IPOPT in a critical real-time application in which I use the CPU time limit option of IPOPT.
> I set the time limit to 475 milliseconds and most of the times the solving process is interrupted at that time (or it finds a solution before that time). However in one particular optimization run it exceeded the limit by 130 milliseconds. Does anyone have an explanation for this?
> I am using the C-interface. IPOPT 3.9.
> Best regards,
> -Martijn
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt

More information about the Ipopt mailing list