Copyright 2020 Ben Eater
This code is MIT licensed.
This is a simple demonstration of the boids algorithm that's featured in this Smarter Every Day video:
This simulation is running on my website if you’d like to check it out.
Each of the boids (bird-oid objects) obeys three simple rules:
Each boid flies towards the the other boids. But they don't just immediately fly directly at each other. They gradually steer towards each other at a rate that you can adjust with the
centeringFactorvariable. In the demo, you can adjust this from 0 to 0.01 with the "coherence" slider.
Each boid also tries to avoid running into the other boids. If it gets too close to another boid it will steer away from it. You can control how quickly it steers with the
avoidFactorvariable. In the demo, you can adjust this from 0 to 0.1 with the "separation" slider.
Finally, each boid tries to match the vector (speed and direction) of the other boids around it. Again, you can control how quickly they try to match vectors using the
centeringFactorvariable. In the demo, you can adjust this from 0 to 0.1 with the "coherence" slider.
There are a ton of ways to extend this simple model to better simulate the behavior of different animals. An example I showed in the video is to limit the "visual range" of each boid. Real animals can't see the entire flock; they can only see the other animals around them. By adjusting the
visualRangevariable, you can adjust how far each boid can "see"—that is which other boids it considers when applying the three rules above.
It ought to run in any web browser. Download (or clone) the files. Then, just double-clicking on
index.htmlon most computers will open the simulation in your web browser. You can then edit
boids.jsto tweak and experiment with the algorithm. Simply save your changes and reload the web browser page to see the effect.
There are lots of features you could try adding to the code yourself:
See this link for more ideas and hints on how to do some of the ideas above.