Oh, now we get the real story! ;-)
If you really want to model a *set* of integers, then either use finite
set constraints (they're done for that), or represent the set as an
*ordered* list of integers.
With finite sets, you can match a list of FD vars with the elements of
the set as:
S={FS.var.upperBound 1#(2*N)} % S is a subset of {1,...,2N}
{FS.card S}=N % S has N elements
Es={FD.list N 1#(2*N)} % Es is a list of ints
{FS.int.match S Es} % Es are the elements of S in ascending order
With finite integers only, just impose a strict ordering:
Es={FD.list N 1#(2*N)}
for X in Es Y in Es.2 do X<:Y end % order all elements
The cost can be obtained by reification. Compute a binary integer that
is equal to 1 iff two elements are consecutive. Then sum them (with
FD.sum of course).
Cost={FD.decl}
Bs = for X in Es Y in Es.2 collect:Collect do
{Collect (X+1 =: Y)} % reify constraint
end
{FD.sum Bs '=:' Cost} % the cost is the sum of the B in Bs
I hope this may help you.
Cheers,
raph
mostafa eslahi wrote:
I'm sorry, there was a problem,
let's me describe again:
I have a alldiff{x(1),..,x(n)} constraint that domain of variables is
[1,..,2n],
This constraint for example (for n=10) produce sets like this {2,5,8,9,1}
now, In the cost function I want to count consecutive values, I mean
cost({2,5,8,9,1})=2 (since 1,2 and 8,9 are 2 consecutive integer).
Now, I hope you get my problem.
Cheers,
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
