[Ipopt] How does IPOPT handle "almost"-zeros in gradients?

Stefan Vigerske svigerske at gams.com
Sun Oct 20 14:05:01 EDT 2019


I don't think that Ipopt treats gradient entries below a certain 
threshold as zero. At least I don't see a need for it and don't recall 
an option that would allow to adjust this.


On 3/4/19 2:36 PM, Simon Helling wrote:
> Dear IPOPT-users,
> I interfaced IPOPT with CasADi in Matlab and unfortunately I think in the gradients provided by CasADi there are many almost-zeros, e.g. 1e-12.
> Does IPOPT ignore values beneath a certain threshold? Would setting these values (which I can’t with CasADi) speed up my computation?
> Thank you in advance,
> Simon Helling
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> https://list.coin-or.org/mailman/listinfo/ipopt

More information about the Ipopt mailing list