Dear all,
we are using the PPL (C++ version) to model a parametric real-time scheduling problem. At some point in our methodology, we obtain a region of space constrained by a C_Polyhedron that describes a subset of valid solutions for our problem.
Our problem has also this nice property: if X=(x_1, ..., x_n) is a vector contained in the C_Polyhedron, then any Y = (y_1, ...., y_n) with
forall i 0<= y_i <= x_i,
is also a valid solution.
Therefore we would like to extend the C_Polyhedron to include all vectors 0<= Y <= X.
So, our questions are:
1) Is there any simple way of doing it in PPL?
In 2D we thought of the following simple strategy: add points (0,0) (x1,0) (0, x2) to the set of generators of the polyhedron, then minimize it.
2) Is this strategy correct? 3) How can we extend it to N dimension?
Thanks in advance for your kind response
Giuseppe Lipari Laboratoire Spécification et Vérification, Ecole Normale Supérieure de Cachan