Re: [eclipse-clp-users] eclipse problem

paniz alipour wrote:
> Hi to all
> bin_packing(Bins, Weights, Capacity) :-
>         dim(Bins,[N]),
>         dim(Weights,[N]),
>         (for(B,1,N),
>          param(Weights,Bins,N,Capacity) do
>              (for(J,1,N),
>              fromto(0,In,Out,Sum),
>              param(Weights,Bins,B) do
> 
>                   Bins[J] #= B
>              ->
>               Out #= In + Weights[J]
>              ;
>               Out = In
>              ),
>              Sum #=< Capacity
>         ).
...
> 
> this is the bin packing problem ,I have a question about bin_packing
> part , about Bin[j]#=B
> this is for what


This seems to be cut from
http://www.hakank.org/eclipse/bin_packing2.ecl
one of the models from Hakan Kjellerstrand's great ECLiPSe web page
at http://www.hakank.org/eclipse/

The intention of this code is to set up the capacity constraint for
a bin, by adding up the weigths of all objects J that are being
put into bin B - that's what Bins[J] #= B means.

This code is actually wrong, because you cannot use a constraint inside
the condition of an if-then-else.  The other variants of the predicate
that Hakan gives in the file are fine - I recommend the last one:

bin_packing4(Bins, Weights, Capacity) :-
        dim(Bins,[N]),
        dim(Weights,[N]),

        ( for(B,1,N),
          param(Weights,Bins,N,Capacity) do
              (for(J,1,N),
               foreach(S,Sum),
               param(Weights,Bins,B) do
                   S = Weights[J]*(Bins[J] #= B)
              ),
              sum(Sum) #=< Capacity
        ).

Here, (Bins[J] #= B) evaluates to 1 if J is in bin B, and 0 otherwise,
so weight[J] gets added only if J is in the bin.



-- Joachim