[Gecode] Announcement: FS constraints prototype

[Christian Schulte]

Currently I have no time until I'm back from travelling in ten days.

Christian

-- 
Christian Schulte, http://www.imit.kth.se/~schulte/ 

> -----Original Message-----
> From: gecode-bounces at ps.uni-sb.de 
> [mailto:gecode-bounces at ps.uni-sb.de] On Behalf Of duchier at ps.uni-sb.de
> Sent: Friday, December 05, 2003 12:03
> To: gecode at ps.uni-sb.de
> Subject: Re: [Gecode] Announcement: FS constraints prototype
> 
> 
> Guido Tack <tack at ps.uni-sb.de> writes:
> 
> >    - S_1\supseteq S_2\cap S_3
> >    - S_1\subseteq S_2\cup S_3
> 
> er... isn't that backwards? it should be:
> 
>       S_1\subseteq S_2\cap S_3
>       S_1\supseteq S_2\cup S_3
> 
> > The model that we think of for cardinality at the moment is the 
> > following: We
> > want to introduce an FD variable together with every FS 
> variable and post a 
> > propagator that is subscribed to both and implements the 
> cardinality 
> > constraints.
> 
> That's essentially the way it is done in Mozart: each FS 
> variable keeps track of min/max for cardinality. FS.card 
> simply posts a propagator between the FS variable and a FD 
> variable that propagates these bounds both ways.
> 
> Please, keep a similar design.  We certainly don't want to 
> introduce cardinality variables and propagators for _all_ sets.
> 
> > The reason for separating the cardinality and the set variable
> > is the limited number of modification events: We just could 
> not signal a 
> > change of cardinality in addition to the MEs for lower and upper
> > bound.
> 
> Could you elaborate on this paragraph?  I am not as familiar 
> with gecode internals as you are :-(
> 
> Cheers,
> 
> -- 
> Denys Duchier - Équipe Calligramme - LORIA, Nancy, France
> 
> _______________________________________________
> Gecode mailing list
> Gecode at ps.uni-sb.de http://www.ps.uni-sb.de/mailman/listinfo/gecode
>