So after Lloyd’s relaxation and drawing Voronoi polygons instead of inhibitory centers it looks pretty good. Here are some samples.
Results look pretty similar to previously when I was using the dart-throwing for the initial positions. Lloyd’s algorithm has way more influence on the positions of the inhibitory centers than the algorithm used to initially place them. Makes sense.
Interestingly, the algorithm is now deterministic, when I regenerate the cross veins I get the exact same output. I’d like to have a little variation so I might add that back in.
Variation could come from the number of inhibitory centers and from their initial placement. I can randomize the number of centers and add a random jitter to their initial positions.
Now I’d like to see how it changes as I adjust that ratio of surface area to the number of inhibitory centers. Slider.
One more pair of wings just because.
