algorithmic modeling for Rhino
All-
Anyone have a good thought on categorizing / grouping wall panels based on a panel dimensions -
I have approached it from two ways - matching all panels to a series of know sizes and looking at an individual panel and finding its match. Both give results but I think they are over complicating the issue.
I think what I should be doing is categorizing? or creating types based on dimensional similarity and outputting the number of variations...? This may be similar to the Morphosis Phare Tower panelization / typing...
Any thoughts would be appreciated.
Anthony
Tags:
thanks for the link- what if you didn't know the number of clusters just that it should be minimized based on a tolerance - and it would be searching keys among multiple dimensional criteria...
I think you can plug Galapagos in the equation, or just a serie of clusters numbers and keep the best one.
About the multi-dimensional criteria, well I dont know, are you talking about (length, width) ?
I think it should be doable to apply k-means on the length, then reapply it on width and get a useful (even if not correct) classification.
I've been a fan of the sort list component. Heres a definition that sorts random placed objects (left side) areas and sorts them along lines neatly by area (right side) using the sort list than delete duplicate and moving with a cluster I made found here. http://www.grasshopper3d.com/forum/topics/move-object-based-on-refe.... Hope it helps, and you can sort anything this way, by color, volume, distance, dimensions. Just change the area list to the desired sorting list.
Welcome to
Grasshopper
© 2025 Created by Scott Davidson.
Powered by