Simulating the Collective Movement of Fish Schools

Abstract

Fish schools are an example of emergent behaviour, where the complexity of the group’s behaviour arises from the simple interaction of the individuals in the group. The emergent behaviour of fish schools is typically simulated using the Boids algorithm. The algorithm is based on three rules that the fish follow: separation, alignment and cohesion. In this thesis, the common method of implementing these rules is discussed and some improvements are proposed. Furthermore, additional rules to the aforementioned three are added to the algorithm. These rules are wander, predator avoidance and obstacle avoidance. The focus of this thesis is developing the algorithm for a visually interesting and lifelike fish schooling simulation. In addition to the algorithm, a demo application is built. The demo application visualizes the created algorithm on a fish school. At the end of the thesis, the visual results are analyzed and further improvements are proposed.