Flocking with Only Two Parameters


  • Dashiell Bhattacharyya Aspiring Scientists’ Summer Internship Program Intern
  • Dr. William Kennedy Aspiring Scientists’ Summer Internship Program Primary Mentor




The boids flocking model, an agent-based model simulating bird flocking behavior, relies on four parameters: the rates of alignment, separation, cohesion, and the vision radius of the birds. Only certain combinations of these parameters result in the birds flocking together. Reliance upon fewer parameters would result in a more efficient and faster modeling process along with clarifying the concepts behind the model's behavior. It was hypothesized that separation and cohesion could be combined into a single parameter and the vision radius removed without affecting the resulting behavior. To test this hypothesis, these simplified models were created, run, and the result of every combination of parameters was distilled to a set of numeric metrics. It was shown that a simplified model combining cohesion and separation to their ratio, termed friendliness, as well as fixing vision radius, was able to result in a functionally equivalent model, demonstrating that the two-parameter model is a successful substitute for the original four-parameter one.





College of Science: Department of Computational and Data Sciences