Beatiful solution, Raphael! Its nicest feature is that it can easily be
changed for the case where Goal is also not determined.
Cheers,
Jorge.
Raphael Collet escreveu:
> Dear Kilian,
>
> I scratched my head quite a bit yesterday, and I just found a nice
> solution to your problem. The idea is to avoid reification, because
> propagation is too poor in that case.
>
> Kilian Sprotte wrote:
>> I am looking for a way to constrain a sequence to be "ascending towards
>> a goal". The goal is determined before posting, let's say its 5, so a
>> solution could be:
>>
>> [1 2 4 5 5]
>
> Let us call this sequence S. It can be defined as:
>
> declare
> Goal=5
> N=5
> S={FD.list N 1#Goal}
>
> The issue is that S is strictly increasing, except when values reach the
> goal. My idea is to combine the propagators <: and FD.min. Consider
> another sequence, say T, that is strictly increasing, and where values
> may be greater than the goal:
>
> declare
> T={FD.list N 1#(Goal+N-1)}
> for A in T B in T.2 do A <: B end % strictly increasing
>
> The link with S is done by taking the minimum between the Goal and each
> element of T:
>
> for X in S Y in T do X={FD.min Goal Y} end
>
> Both propagators <: and FD.min will propagate of the variables' domains,
> and propagates exactly as you wanted:
>
>> [1#5 2#5 3#5 4#5 5]
>
> Try to constrain the 3rd element of S to value 4 in the example, and you
> will see that the 4th element is constrained to 5.
>
>
> Cheers,
> raph
> _________________________________________________________________________________
> mozart-users mailing list
> [email protected]
> http://www.mozart-oz.org/mailman/listinfo/mozart-users
>
Jorge M. Pelizzoni
ICMC - Universidade de São Paulo
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
