[SciPy-user] nonlinear optimisation with constraints
Tue Jun 23 03:00:31 CDT 2009
On Mon, Jun 22, 2009 at 7:20 PM, Robert Kern<email@example.com> wrote:
> 2009/6/22 Ernest Adrogué <firstname.lastname@example.org>:
>> Mmmm, yes, but the box constraints are merely to prevent the
>> algorithm from evaluating f(x) with values of x for which f(x)
>> is not defined. It's not a "real" constraint, because I know
>> beforehand that all elements of x are > 0 at the maximum.
> Unfortunately, that's not how constraints work in most optimizers.
> Usually, the infeasible region is necessarily also explored.
Do you have a certain solver in mind?
>From what I read in the literature, I thought that for simple box
constraints usually some projection to the set of feasible search
directions is performed.
And only for the nonlinear constraints one uses a merit function and
an active set strategy for the inequality constraints which may yield
infeasible steps in the process.
> Robert Kern
> "I have come to believe that the whole world is an enigma, a harmless
> enigma that is made terrible by our own mad attempt to interpret it as
> though it had an underlying truth."
> -- Umberto Eco
> SciPy-user mailing list
More information about the SciPy-user