[gecode-users] Space entailment

[Christian Schulte]

Hi,

well, yes and no. The difference between succeeded(entailed) and
succeeded(stuck) is easy to find: just check the number of propagators left.


The other one does not work: if you post a new propagator it might find out
that it is entailed and disappear. However, it might also propagate and some
other propagator becomes entailed. So you don't know whether the newly
posted propagator is actually entailed.

Maybe you tell us what you want to do exactly?

Christian

-----Original Message-----
From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Gustavo Gutierrez
Sent: Thursday, March 01, 2007 1:43 PM
To: gecode at gecode.org
Subject: [gecode-users] Space entailment


Hi all,

I was wondering about the representation of the entailment relation in
gecode. Suppose you have a space with variables and constraints. How can i
verify if an extra constraint is already entailed by the store (space)?. In
the mozart terminology this is the difference between a
succeeded(entailed) and a succeeded(stuck) space. Of course, here we will
not take threads into account since threads does not exist in gecode.

At implementation level, if the relation implemented by a propagator to be
posted in the space is already represented in terms of variables domains,
the propagator is not posted at all. Then counting how many propagators are
in the space before the posting of the new one and comparing it to the
number of propagators after posting it, is enough to verify that property.

Do you agree??


Cheers,
--
Gustavo Gutierrez

_______________________________________________
Gecode users mailing list
users at gecode.org https://www.gecode.org/mailman/listinfo/gecode-users