output: github_document

boids4R, Deterministic Reynolds-style boids and swarm simulations boids4R logo

Frédéric Bertrand

boids4R provides deterministic Reynolds-style boids and swarm simulations for R. It owns simulation semantics only: agents, rules, obstacles, attractors, predators, species, and frame export. Rendering packages consume the exported frames through optional adapters.

The package is intended to pair cleanly with browser-native renderers such as ggWebGL, without putting camera, shader, widget, or scene fields into the core simulation objects.

Example

library(boids4R)

sim <- boids_scenario("murmuration_3d", n = 400, steps = 240, seed = 1)
frames <- as.data.frame(sim)
head(frames)

if (requireNamespace("ggWebGL", quietly = TRUE) &&
    utils::packageVersion("ggWebGL") >= "0.4.0") {
  spec <- as_ggwebgl_spec(sim, trail_length = 30)
  ggWebGL::ggWebGL(spec, height = 520)
}

The WebGL adapter emphasizes current boids, keeps recent history as faint trails, colours species and explicit prey/predator roles distinctly, and draws velocity arrows in species colours by default.

References

The model follows the standard boids lineage introduced by Craig Reynolds: separation, alignment, and cohesion rules produce flocking behavior through local agent interactions.