A Greedy Flock: How Foraging Gives Rise to Collective Behavior ✛

Original: "Emergent aggregation from collective foraging"
arXiv:2608.28046v1 · 2026-08-28 · CC BY-SA 4.0 · 2 min
Agents trained to find targets without any direct reward for grouping spontaneously form clusters.

Picture a flock of birds gathering around a feeder that no one built. Each bird searches for food on its own, yet they all end up together. Usually, such behavior is explained by social instinct: birds want to be near their kin. But what if the flock is simply a side effect of greed?

The researchers trained 50 agents to find targets without seeing the targets themselves, only by observing other agents. Reward was given solely for capturing a target.

The agents lived on a square arena with sides of 50 steps. Scattered around were 100 "feeders"—targets that, once captured, disappeared for a while. Each agent could see only what lay ahead, within a narrow cone of vision. It could tell whether a conspecific was there and whether that individual had recently visited a feeder. There was no reward for being near others—only for food.

With a small visual radius, the agents behaved like loners: they wandered across the arena, occasionally bumping into targets. But when the visual radius was increased, a sharp jump occurred: the agents began to spin in place until they spotted a "satiated" conspecific, then ran straight toward it. At that moment, they clumped together.

The transition was sharp but not phase-like: it is a crossover, not a thermodynamic transition.

This transition can be explained by a simple model. The solitary strategy is blind search, like a person looking for keys in the dark. The collective strategy is running toward someone who has already found the prize. When the visual radius is small, it pays to search on your own. When it is large, it pays to watch others. Somewhere in between lies the point where the benefits balance out.

So a flock can arise from pure greed. Agents do not need to want to be together—they simply need to want to eat. And this finding changes how we view collective behavior: perhaps birds and fish school not because they enjoy company, but because it is the best way to find dinner.

The Clark–Evans index, used here to quantify aggregation, was proposed in 1954 by ecologists to analyze the spatial distribution of plants and animals.

The idea that complex collective behavior can emerge from simple individual rules without central control echoes the concept of the 'swarm' in Stanisław Lem's novel 'The Invincible', where microscopic machines form destructive structures.

\lambda^* = \frac{\ell}{W\left(\frac{\ell e^{K}}{\tau}\right)}
λ* is the critical detection radius, ℓ is the reactive length, τ is the target depletion time, W is the Lambert W-function, and K is a constant depending on target radius and density.
Concepts
emergent collective learning collective behavior Reinforcement Learning founder effect agent random forest emergent complexity modeling complex systems active matter random network ensemble average positive feedback loop Effective field theory entropy driven self assembly evolution epidemic optimization ergodic theorem principle of least action clustering