[Help-gsl] Re: over-constrained multidimensional root finding

[Brian Hawkins]

Could you reform the problem into a minimization problem? (e.g., solve
min(y**2) instead of y==0)  That might offer more flexibility.


It appears that all of GSL's multidimensional root finding routines
> require that the number of constraints equals the number of variables.
> For the problems that I want to solve, this rarely is the case.
> Typically I have constraints that outnumber the variables, and it may or
> may not be the case that some of these end up being redundant (it is not
> easy for me to determine this) so the problem may be over-constrained or
> under-constrained.  Does GSL have any root finders that would be
> applicable in this case?  Currently I rely on the annealing algorithms
> but this does not find the solution very fast.
>
> Thanks,
> Dan
>