Re: Limits of ECLiPSe ?

From: Warwick Harvey <wh_at_icparc.ic.ac.uk>
Date: Mon 22 Oct 2001 03:19:00 PM GMT
Message-ID: <20011022161900.K22153@tempest.icparc.ic.ac.uk>
On Mon, Oct 22, 2001 at 02:57:13PM +0000, Olivier Thirifay wrote:
> I would like to known the limits of ECLiPSe with domain (maximum 
> variable,...)
> 
> I have written a clause :
> 
> predicat(VARS) :-
> VARS=[A,B,C,D],
> A::0..1000,
> B::0..1000,
> C::0..1000,
> D::0..1000,
> A#>499,B#>2,C#>200,D#>1,
> labeling(VARS).
> 
> ECLiPSe crash with it ! I suppose there are too much possibilities for it.
> If I reduce the domain or change an inequality in an equality, it runs..
> My problem is that I have to use more variables and maybe wider domains.
> Can you help me to resolve this problem or my problem is too big for ECLiPSe 
> ?

I doubt you mean that ECLiPSe crashed.  If it did, please provide us with
more information, such as the error message reported, the exact query used,
which version of ECLiPSe you were using, and which operating system you were
using it with.

What you probably mean is that you ran out of patience before ECLiPSe had
explored all the possibilities.  There are 399598401600 different solutions
to your predicate.  Even checking a million of them every second would take
more than 4 days.  But probably you don't want to go through every
combination, so you need to add constraints which limit the search to the
interesting solutions before you call the labelling predicate.  If this is
done right, ECLiPSe will need to explore only a small fraction of this huge
solution space, and will probably return you the answer you want in a much
more reasonable time.

Cheers,
Warwick
Received on Mon Oct 22 16:19:34 2001

This archive was generated by hypermail 2.1.8 : Wed 16 Nov 2005 06:07:10 PM GMT GMT