[Ipopt] Lagrange Multipliers, Scaling, and convergence tolerance

[Stefan Vigerske]

Hi,

I never used the user-scaling option, but I would presume that the 
returned multipliers are not scaled, that is, that they correspond to 
the original problem. That's how it is for the other scaling options and 
I don't think it should be different when using user defined scaling values.

You can just try it out by looking at the multipliers with your 
user-scaling enabled and disabled.

Best,
Stefan

On 02/28/2017 11:26 PM, Brad Bell wrote:
> I am trying to understand the definition of the Lagrange multiples
> lambda returned by
> finalize_solution when the scaling method is user-scaling.
>
> It is not clear to me (from the documentation) if these multipliers are
> for the the original space, where f(x) and g(x) are defined, or in the
> scaled problem ?
>
> I am using the Lagrange multipliers to check the convergence tolerance
> which
> seems to me (form the documentation) to be in the scaled space.
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> https://urldefense.proofpoint.com/v2/url?u=http-3A__list.coin-2Dor.org_mailman_listinfo_ipopt&d=DwICAg&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=BRcuJnQr5NAzU29t80hk2rsLc4vrlRySBDabuq0O1ZI&m=HihXZmEg_vZjBizuwu0W5rjHSOXVU5-fsTigUDiVel4&s=voAZWaVtzCsTcwWMqb0U6LidhII5smNYkQAGOL9b0NM&e=
>


-- 
http://www.gams.com/~stefan