[gecode-users] multi-dimensional bin-packing constraint - bin capacities

[Christian Schulte]

Hi Kish,

Yes, when I added this I thought exactly the same. However, I decided to
include the c argument to avoid confusion: the values of c (or as you say
the upper bound on the load variables) actually defines how well the
constraint is propagated. So what I tried to avoid by including c is that
users first post the binpacking constraint and only then adjust the load
variables...

I know, matter of taste.

Cheers
Christian

--
Christian Schulte, Professor of Computer Science, KTH,
www.gecode.org/~schulte/


-----Original Message-----
From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Kish Shen
Sent: Monday, November 17, 2014 6:52 PM
To: users at gecode.org
Subject: [gecode-users] multi-dimensional bin-packing constraint - bin
capacities

Hi,

I am updating ECLiPSe's Gecode interface to Gecode 4.3, and I am adding
support for the multi-dimensional bin-packing constraint.

I have a small question about the arguments for the constraint:

binpacking(home, d, l, b, s, c)

has the capacities argument c that is not found in the 1-D bin-packing
constraint:

binpacking(home, l, b, s)

My understanding is that for this form of the constraint, c is implicitly
specified by l, the loads, i.e. the domains of each bin is specified by the
domain in the load variable for that bin.

It seems to me that the values for c is also implicitly specified for the
multi-dimensional case, and can be derived from l, as the maximum domain
value of all the variables in l for each dimension. I am thinking of
providing the constraint in this form so that it looks more like the 1-D
version. Does this make sense?

Cheers,

Kish




_______________________________________________
Gecode users mailing list
users at gecode.org
https://www.gecode.org/mailman/listinfo/gecode-users