Ant Colony Double-Bridge Experiment

Bridge Setup

Long / short path length ratio 2.0×

Pheromone evaporation rate 0.12 /s

Choice nonlinearity n 2.0

Colony size 50 ants

Live State

short 50%long 50%

Pheromone (short)

0.0

Pheromone (long)

0.0

Round trips completed

How it works

This reconstructs the asymmetric double-bridge experiment (Deneubourg, Aron, Goss & Pasteels, 1990): a colony connects nest and food source through two branches of different length. Ants lay no directional signal at the fork — they simply deposit pheromone along whichever branch they just walked, and pheromone evaporates continuously. At each decision point an ant chooses a branch with probability weighted by the pheromone already on it:

P(short) = (c + φ_short)^n
           ───────────────────────────────
           (c + φ_short)^n + (c + φ_long)^n

φ evaporates:  φ(t+dt) = φ(t) · e^(−ρ·dt)
φ on arrival:  φ ← φ + Δ  (one ant's deposit)

c is a small attraction threshold (kept fixed here) that keeps early choices close to 50/50 before any trail exists; n controls how strongly ants amplify a pheromone lead once one appears.

Length ratio — how much longer the long branch is. Ants that pick the short branch complete round trips faster, so they reinforce its pheromone more often per unit time — pure positive feedback with no ant ever comparing the two lengths directly.
Evaporation rate ρ — higher ρ erases old choices faster, keeping the colony flexible but slower to lock in; ρ → 0 makes any early random majority permanent.
Nonlinearity n — n = 1 gives proportional (linear) probability matching; n > 1, as measured for real Argentine ants (Linepithema humile), makes a small pheromone lead snowball into near-total convergence.
Colony size — more ants sample the fork more often per second, so the network converges faster relative to evaporation.

This mechanism — stigmergic reinforcement of shorter paths through nothing but evaporating pheromone and a probabilistic choice rule — is the exact biological basis Marco Dorigo formalised in 1992 as Ant Colony Optimization, the metaheuristic used today for routing and combinatorial search.

Short-branch pheromone Long-branch pheromone Ants

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.