Re: qp() issues?

[fork]

Juan Pablo Carbajal <carbajal <at> ifi.uzh.ch> writes:

> then you can use equality constraints to enforce the difference. That
> is if x is the current value of the  vector variable (of length n) you
> want
> 
> x(2:n) - x(1:n-1) + dx = 0
> 
> I hope this helps

Well, it is definitely cool.  I am going to keep hacking at the formulation
using inequality matrixes in qp(), and maybe a simple test formulation, because
either there is a bug (unlikely but possible), or I don't understand something
(more likely).  

I may wind up rewriting the constraints and trying to call __qp__ directly, and
slack ish techniques like the above will really come in handy.

Tx