[gecode-users] Access to the last propagator invoked in case of ES_FAILED
George Katsirelos
gkatsi at gmail.com
Thu Mar 27 15:41:13 CET 2008
Hello,
I also needed that information to implement clause learning. All I did
was modify Space::propagate() to keep track of the propagator that
failed. I admit it may not cover all cases (for example it does cover
failure when the propagator was called from Space::step()), but during
search this seems to be the only path that leads to failure.
George
On 28/03/2008, Christian Schulte <cschulte at kth.se> wrote:
> Actually, I have to say that this is not that easy: failure might actually
> not be due to the propagator run last. It could be due to a variable
> modified outside a propagator (when posting constraint or using a
> branching).
>
> Moreover, the propagator run last is typically not the culprit of things
> going wrong.
>
> Cheers
> Christian
>
>
> --
> Christian Schulte, www.ict.kth.se/~cschulte/
>
>
>
> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
> Of Mikael Zayenz Lagerkvist
> Sent: Thursday, March 27, 2008 3:27 PM
> To: David Rijsman
> Cc: users at gecode.org
> Subject: Re: [gecode-users] Access to the last propagator invoked in case of
> ES_FAILED
>
> Hi,
>
> There is no way currently to access this information, and no current
> plans to introduce it. However, it would not be hard to implement
> yourself. As a starting-point, see the implementation of the function
> Space::status in gecode/kernel/core.cc.
>
> What is it that you want to do with this information? Since
> propagators do not have any identifier that is stable between copies,
> the information seems to me to be of little use.
>
> Cheers,
> Mikael
>
> On Thu, Mar 27, 2008 at 3:03 PM, David Rijsman <davidrijsman at gmail.com>
> wrote:
> > I would like to know if it is possible to determine the last propator
> > invoked in case Space::status returns ES_FAILED or if there are any
> > plans to expose this information. If one can get this information (in
> > the future) will one also introduce reflection API on the Propagator
> > level?
> >
> > David Rijsman
> >
> > _______________________________________________
> > Gecode users mailing list
> > users at gecode.org
> > https://www.gecode.org/mailman/listinfo/gecode-users
> >
>
>
>
> --
> Mikael Zayenz Lagerkvist, http://www.ict.kth.se/~zayenz/
>
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
>
>
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
>
More information about the gecode-users
mailing list