Ant Colony Pheromone Foraging (2D)

Colony Parameters

Number of ants 150

Pheromone evaporation 15%/s

Click the stage to drop a food source, then watch the trails form.

Live Statistics

Foraging / returning

0 / 0

Food delivered

Active food sources

Path efficiency

–

How it works

This is a real two-pheromone ant colony optimization (ACO) model, run top-down on a grid. Every ant is in one of two states: searching (leaving the nest, laying a decaying "to-home" trail as it goes, and steering up the "to-food" gradient it senses ahead) or returning (carrying food back, laying a "to-food" trail, and steering up the "to-home" gradient). Both pheromone fields evaporate every frame at the slider's rate, so trails that no one reinforces fade — only paths that many ants keep re-walking survive, which is exactly the stigmergic feedback loop that lets a colony with no map converge on the shortest route.

Number of ants — more ants reinforce a trail faster and find shorter alternate routes sooner.
Pheromone evaporation — higher rates prune stale trails quickly (faster re-routing around a depleted food source, but noisier early exploration); lower rates let a trail persist even if it isn't optimal.
Place food / Place obstacle — click the stage to drop a food source (ants deplete it as they carry pieces home) or a wall (ants steer around it, which lets you watch a trail detour and then straighten out as pheromone reshapes the path).
Path efficiency — the running average of each completed round trip's step count divided by the straight-line distance to the food it visited; it starts high (random exploration) and drops toward 1.0 as the pheromone trail straightens toward the shortest walkable path.

Click: place · drag also works

To-home pheromone To-food pheromone Obstacle Food source