[Ipopt] Initial function evaluation not at initial point

Stefan Vigerske stefan at math.hu-berlin.de
Thu Apr 19 22:27:16 EDT 2012


Hi,

Ipopt moves the starting point away from the variable bounds, if it is 
too close, see options bound_frac and bound_push:
http://www.coin-or.org/Ipopt/documentation/node97.html
http://www.coin-or.org/Ipopt/documentation/node98.html

Could that be the cause that you see you initial point moved?

Stefan

> Hi,
>
> When I start my problem from a trivial initial start value, all
> functions are evaluated at the initial design.
>
> When I start from a design (obtained by another optimizer) which is
> optimal and partially close to the design bounds,
> the function gradients are evaluated at the initial design but the
> function evaluations are evaluated at another design.
>
> If I relax the bounds of the design, the functions are evaluated at a
> slightly different other design.
>
> Can this be? I see not much chance that this is caused by a bug by
> myself, but one can never be sure.
>
> Thanks a lot,
>
> Fabian
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>



More information about the Ipopt mailing list