Packing/nesting rectangle solutions - generative plans

Hello everyone,

From what I've seen that's an old questions, I've found post starting on 2010 ! And after viewing 50+ discussions related on the forum, I can't figure out what's the best one.

So my problem is I want to pack squares/rectangles (so all in 2D) with constraints (boundaries, overlaping or no). The idea is to get as many differents attempts as possible and to compare them, so optimization, but not too much.

So far I've tried it with :
- Galapagos (seems to be the best one so far)
with Tuğrul Yazar method.
- PackRat (good but not enough to tend to "optimization")
- KDTree solutions for solving generative plans
- Space Syntax from Pirouz Nourian (http://www.grasshopper3d.com/group/space-syntax) maybe too specific to architecture, I'd like a more geometrical and open solution.

I've also review the work of Nate Holland & Martin Stacey ( and Jesse attempt to redo it http://www.grasshopper3d.com/forum/topics/realtime-physics-for-spac...).

And for those interested in generative plans, I advise you the thesis of Ron Aasholm "Incessant Replication - Computational Floor Plan Generation".

So I'm gonna make a 2D vision of what I have in mind in a grasshopper way :
I want a way to organise/nest/pack defined rectilign closed curves in 2D according to constraints, and having differents solutions.

Anyway this subject is really interesting, I'll post updates here, and if you can point me the right direction it would be great !

I'm not posting any .gh file at this point, that would be only upgraded or modified versions of the references I'm speaking in my post.