[gecode-users] constraint between variables in different spaces

[Christian Schulte]

No, cloning is never done implicitly.

 

Christian

 

--

Christian Schulte, www.ict.kth.se/~cschulte/

 

From: Mauricio Toro [mailto:mauriciotorob at gmail.com] 
Sent: Wednesday, May 21, 2008 1:14 PM
To: cschulte at kth.se; users at gecode.org
Subject: Re: constraint between variables in different spaces

 

If I am not using search/branching should I worry about cloning spaces?

Does the spaces clone in other cases different than when I do braching?

 

Thanks

2008/5/21 Christian Schulte <cschulte at kth.se>:

Well that would not be channeling, it would just transfer the current domain
from one space to another. As said, it can be done but it is non-trivial: it
would require locking as different spaces can be run by different threads.

 

What I meant by copying is: suppose that you have a propagator between
spaces s1 and s2. Now you clone s1. What should happen to the propagator?
Should it be cloned as well?

 

If you do not use cloning you could build a propagator, it would basically
be two propagators, one in each space.

 

I am not yet convinced that Gecode is a good platform for CCP, it is really
designed for constraint reasoning.

 

Christian

 

--

Christian Schulte, www.ict.kth.se/~cschulte/

 

From: Mauricio Toro [mailto:mauriciotorob at gmail.com] 
Sent: Wednesday, May 21, 2008 12:09 PM
To: cschulte at kth.se; users at gecode.org
Subject: constraint between variables in different spaces

 



Hello all,

 

Is it possible to do the channeling for IntVar using a IntSet?

like dom(space1,var1,intset); dom(space2,var2,intset); ?

 

Why do you mean by copying it?

 

I have made an interpreter for CCP calculus using gecode.

Now I am trying to extend the interpreter for Time Concurrent Constraint
Calculus.

which is an extension where we have multiple spaces.

each one representing a time unit. The problem is when we want to represent 

persistent assignation, something like "forever tell(x =
Variable_In_Time_Unit_1)".

 

Notice that for all this models we do not use search, we are only interested
in propagation.

 

Any ideas?

 

Thanks

 

2008/5/21 Christian Schulte <cschulte at kth.se>:

 

You can't. The best you might is a segfault. There is currently no way to do
it. One could build a special propagator that channels across spaces but
that would be pretty tricky: who will take care of that propagator during
copying, for example.

 

My question is: why would you like to do that?

 

Cheers

Christian

 

--

Christian Schulte, www.ict.kth.se/~cschulte/

 

From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Mauricio Toro
Sent: Wednesday, May 21, 2008 10:39 AM
To: users at gecode.org
Subject: [gecode-users] constraint between variables in different spaces

 

Hello all,

I want to propagate a constraint between variables in different spaces.

How can this be done?

 

For example

 

mySpace * a = new mySpace();

mySpace * b = new mySpace();

IntVar aa(a,1,10);

IntVar bb(b,1,10);

rel(a,aa,IRT_EQ,2);

a->status();

rel(b,bb,IRT_EQ,aa);

b->status();

std::cout << aa << " " << bb; // should print : 2 2

 

It does not work this way. Is there a way to do it?

Maybe cloning the variable or something?

 

Thanks


-- 
Mauricio Toro Bermudez
Estudiante de Ingeniería de Sistemas
Pontificia Universidad Javeriana, Colombia

Stagiare à l'Ircam
1, place Igor-Stravinsky 75004 Paris, 
France de 2008 à 2009 




-- 
Mauricio Toro Bermudez
Estudiante de Ingeniería de Sistemas
Pontificia Universidad Javeriana, Colombia

Stagiare à l'Ircam
1, place Igor-Stravinsky 75004 Paris, 
France de 2008 à 2009 




-- 
Mauricio Toro Bermudez
Estudiante de Ingeniería de Sistemas
Pontificia Universidad Javeriana, Colombia

Stagiare à l'Ircam
1, place Igor-Stravinsky 75004 Paris, 
France de 2008 à 2009 




-- 
Mauricio Toro Bermudez
Estudiante de Ingeniería de Sistemas
Pontificia Universidad Javeriana, Colombia

Stagiare à l'Ircam
1, place Igor-Stravinsky 75004 Paris, 
France de 2008 à 2009 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ps.uni-sb.de/pipermail/users/attachments/20080521/6a420102/attachment.htm>