Emergence of a giant rotating cluster of fish in three dimensions by local interactions

TL;DR

This paper presents an agent-based model explaining how schooling fish form giant rotating clusters through limited local interactions, reproducing observed patterns and scaling laws without long-range interactions.

Contribution

The model introduces interaction capacity constraints and autonomous control to replicate large rotating fish clusters, aligning with experimental scaling laws.

Findings

Reproduces giant rotating clusters like balls, tori, and rings.

Matches the scaling law between agent number and cluster area.

Shows clusters form at low interaction capacity without long-range interactions.

Abstract

Schooling fish exhibit giant rotating clusters such as balls, tori, and rings, among other collective patterns. In order to account for their giantness and flexible shape change, we introduce an agent-based model that limits the number of agents that each agent can interact with (interaction capacity). Incorporating autonomous control of attractive interactions, we reproduce rotating clusters (balls, tori, and rings) that are an order of magnitude larger than the interaction range. We obtained a phase diagram of patterns including polarized schools and swarms. In our model, the scaling law between the number of agents and the projected area of the cluster is in good agreement with experimental results. The model indicates that giant rotating clusters are formed at low interaction capacity, without long-range interactions or inherent chirality of fish.