Re: new function for quadratic programming: pqpnonneg

[Jaroslav Hajek]

On Sat, Oct 31, 2009 at 4:23 PM, Jonathan Stickel <address@hidden> wrote:
> I finally got a round to checking this function out.  However, my problem
> also has equality and inequality constraints in addition to x>=0.  Is it
> possible to add constraints and use the same implementation
>
> Thanks,
> Jonathan
>

I don't think this can be done in full generality. It might be
possible to solve for the constraints first and then project the
quadratic form to the feasible space. If you're lucky and the feasible
space is a convex "corner", you can still use pqpnonneg.
The simple "corner-like" shape of the feasible space is where the
algorithm gets its good features from. A generalization to arbitrary
combination of lower and upper bounds is possible, but hasn't been
implemented yet. I'm planning to address this in the future.

hth


-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz