[rules-users] Drools Planner changing problem fact (2D bin packing with related dimension and continuous values)

[Wolfgang Laun]

I don't know anything about Planner, so pardon me if this isn't helping any.

Allocating rectangles into a bigger rectangle could be discretised,
provided the dimensions have a certain granularity. This means that
you have a gridded "arena", from which you allocate certain grid
subsets for the "cuts".

There are some well known techniques for representing gridded areas,
and the corresponding geospatial functions are quite efficient.

I surmise that Planner constraints can be formulated based on a
gridded approach, but (see 1st sentence) I wouldn't know how.

-W

PS: From some research done ~30 years ago I have a vague recollection
that the 2D cutting stock problem is a well trodden area. Google
produces some newer results as well.



On 20/06/2012, Ralph S. <ralphschwitalla at me.com> wrote:
> Hi Geoffrey,
> thanks for the answer.
> I like the idea with the getPrevious(X/Y)Item from a combinatorial point of
> view, but practially it doesn´t work because you can have more than one
> previous item (bigger one has 2 smaller ones to the left/above) and vice
> versa, which will let the combinations also explode.
>
> I still like to implement an online Guillotine algorithm (cut the remaining
> space and do an heuristic search on the possible results) in drools if it´s
> possible.  Therefore i still have no idea how to handle the dynamically
> changing list of free spaces.
>
> Greetings
> Ralph
>
> --
> View this message in context:
> http://drools.46999.n3.nabble.com/Drools-Planner-changing-problem-fact-tp4018087p4018101.html
> Sent from the Drools: User forum mailing list archive at Nabble.com.
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>