Differential Growth

Comment

Comment by Eva James on December 10, 2017 at 3:11pm

Hi David,

Would you consider sharing the Grasshopper definition of this study http://spatialslur.com/differential-growth-2-2/ ? Thank you. 

Comment by David Reeves on March 25, 2016 at 7:54pm

Thanks guys.

Nik -
I've had some luck constraining/biasing the growth with volumetric information - a scalar field representing signed distance to a boundary mesh for example.

Hans - 
Yeah I think Andy is the patient zero for this one. In any case, feel free to build upon the C# implementation. 

Regarding vertex collisions, a KdTree probably won't give the performance boost you'd expect in this case. KdTrees are quick to search but relatively slow to build (each insertion is O(log n)) and because the point set is dynamic, you'd need to rebuild every iteration. The simulation is better suited to spatial hashing or even a simple uniform grid - especially since collision objects are all the same size.

Comment by Hans Jakob on March 25, 2016 at 11:41am

its pretty much amazing, how this type of algorithms make the rounds.
this is the project that we did in Phyton last Semester at University of Stuttgart. 
http://issuu.com/hansjakobwagner/docs/albahar_giachini_wagner_web/1
https://vimeo.com/156077002

It lacks very much the KD-Tree implementation and so the Maximum Number of interacting cells was limited to smth like 74k, taking literally hours to compute. 

Wanted to make a C# version of it now for speedup-purposes, but stumbled on this post here and was very amazed. 
If its ok for you, ill try to implement some more functions ...

Comment by Jens Pedersen on March 24, 2016 at 8:55am

drop it like its hot Dave! 

Comment by Anders Holden Deleuran on March 24, 2016 at 2:39am

That's hot!

Comment by David Reeves on March 23, 2016 at 8:07pm

Also a shout out to the clever folks involved in this discussion