[PPL-devel] simplifying a constraint system?

[Roberto Bagnara]

Basile STARYNKEVITCH wrote:
> [Sorry for the very naive & basic question]

Hi Basile,

does this mean you did not read the manual?  :-)

> I have constructed some constraint system using the C API. I know I'm 
> done constructing it.

A "constraint system" is a PPL syntactic object.  As such, it is only
one possible (very limited) representation for a PPL semantic object,
which is what you probably want to work with.

> Is there some function to "simplify" it or represent it more efficiently?
> - perhaps onverting it to NNC_Polyhedron and backwards, using 
> ppl_new_NNC_Polyhedron_from_Constraint_System 

Why NNC_Polyhedron?  Do you have strict constraints?  If not you should
use C_Polyhedron.

> & ppl_Polyhedron_get_minimuzed_constraints 
> ?

Yes, this sequence will give you a minimized constraint system
(in the precise sense stated in the documentation).

> Is there some function to test that it is empty?

Of course:

int ppl_Polyhedron_is_empty(ppl_const_Polyhedron_t ph)

 > - perhaps  using [...]

No.

> A concrete case occuring to me is the constraint system of two variables 
> n and y with constraints {n = 10, y = 17, -n >= 0} I need to compute 
> that it is empty.

See above.

> Is there some function to test that it has integer elements, in the 
> sense that there exist an integer setting of the variables such as the 
> constraint are satisfied?

int ppl_Polyhedron_contains_integer_point(ppl_const_Polyhedron_t ph)

Beware though that this operation can be very expensive.

> - perhaps using grids?

No.

> Sorry for such basic questions!

It's OK.  But if you don't read the library documentation I am afraid
you will not go very far.
Cheers,

    Roberto

-- 
Prof. Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara at cs.unipr.it