I think this is a really interesting topic.
I have tried something along these lines once a while back - experimenting with the idea of valence, like in chemical systems.
I think the results were interesting, and quite biological looking, but I didn’t get it to where I was imagining in terms of the structures formed.
Here’s where I got to with some of those tests:
and here’s the file
valence_join_tests.gh (12.9 KB)
Choosing the thresholds for attraction and repulsion is critical.
The other thing that I think might make this work better is some way of stopping them just forming triangles, by making a point’s connected neighbours repel each other. It also might be better to work with line segments from the start.
Also here the attraction and repulsion is spring-like in the way its strength varies with distance, but that introduces a discontinuity (since when points get below the threshold distance the attraction suddenly kicks in at a high strength), and this might be bad for stability. A custom distance based attraction could help.
Anyway, I think there’s a lot more that could be played with here…