[Ipopt] Using IPOPT for systems of nonlinear equations

[Stefan Vigerske]

Hi,

the derivative tester in Ipopt is always a good tool to debug these 
kinds of problems.
I doubt that your gradient is correct. I think it should be
A(1) = 2.*(X(1)-1.)

Stefan

On 11/19/2012 01:56 PM, simon.scheidegger at bf.uzh.ch wrote:
> Hi everyone,
>
> I would like to use IPOPT to solve systems of nonlinear equations
> (as I use it already at another location of the code to solve an optimization problem,
> so  this would be very handy).
>
> However, I somehow fail to set the options for IPOPT appropriately for this purpose:
>
> I was proceeding as follows:
> In order to use an optimizer to solve systems of nonlinear equations,
> this can be achieved by setting the objective function to zero, i.e.
>
>   min  f(x) = 0
>     x
>
> and putting the system of nonlinear equations as "equality constraints".
>
> An easy example:
>
> objective function:
> f(x) = 0
>
> nonlinear equation:
>
> f1 : (x -1)^2 =0
>
>
> However, when I run this example (following the example of the tutorial, i.e. hs071_f.f), i obtain
> wrong results.
>
> I attached the log file, the makefile and the corresponding Fortran example.
>
> I would be glad if someone could help me here.
>
> Kind regards,
>
> Simon
>
>
>
>
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>