[gecode-users] Negative table in Gecode

[Peter Nightingale]

Hi,

Thanks for the suggestions Christian and George (did you intend to post 
to the list George?). For now I have decided I need to make a release 
because the previous release has a fairly serious bug. I have done 
something quick and easy -- decompose into one constraint to rule out 
each tuple.  I should come back to this at some point.

I'm tempted to have a go at porting the MDDC and negative MDDC 
propagators that I wrote for Minion. However Gecode and Minion are 
different enough that I doubt it would be easy.

Pete

On 27/02/15 15:46, Christian Schulte wrote:
> Hi Pete,
>
> Yes, I have a brilliant idea: how about you implement a negative table
> constraint in Gecode! We would love it.
>
> I think you get the gist: no, there is nothing really...
>
> Cheers
> Christian
>
> --
> Christian Schulte, KTH, web.it.kth.se/~cschulte/
>
> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
> Of Peter Nightingale
> Sent: Friday, February 27, 2015 03:11 PM
> To: users at gecode.org
> Subject: [gecode-users] Negative table in Gecode
>
> Hi,
>
> I hope you're all well, and thoroughly enjoying your IJCAI reviewing!
>
> I'm working on Savile Row's Gecode (flatzinc) output -- I have a type in SR
> for negative table constraints (i.e. the tuple set represents the disallowed
> tuples). I don't see anything that looks relevant in registry.cpp.
> Obviously turning it into a positive table constraint is an exponential
> explosion.
>
> I could decompose into one constraint per tuple -- but of course this is
> horrible for propagation.
>
> Can you suggest anything better?
>
> Thanks,
> Peter
>
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
>
>
>