[Ipopt] Problem has too few degrees of freedom

[Stefan Vigerske]

Hi,

yes, Ipopt first checks whether the number of (unfixed) variables is 
greater-or-equal the number of equality constraints. Otherwise, it will 
regard the problem as overconstrained and stop with the message you see.
This conclusions is based on assumptions on constrained qualifications, 
basically, that there are no redundant equality constraints.

Stefan

On 03/25/2015 03:49 PM, Ge, Ming wrote:
> Hi everyone,
>
> I used IPOPT to solve a nonlinear problem and get an optimal solution [cid:image001.png at 01D0670A.F58851B0] . But it goes wrong when I attempted to add another linear constraint [cid:image002.png at 01D0670A.F58851B0]  to the original problem, where [cid:image003.png at 01D0670A.F58851B0]  still holds. The error message is 'Problem has too few degrees of freedom' even the optimal solution [cid:image004.png at 01D0670A.F58851B0]  is provided as initial searching point. I wonder whether this is because IPOPT would check the degrees of freedom first then begin to search? Otherwise it could just end up with the initial searching point provided as it is the solution that satisfies all the constraints. Please advise.
>
> Many thanks
>
>
> Steve
>
>
>
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>