A general perception based framework for modelling animal aggregation

Animal aggregates (sparrow flocks, anchovy schools, caribou herds) are complex systems potentially made up of thousands of individuals moving in a coordinated mass. While such aggregates are difficult to study in the field, individual based spatially explicit computer models (IBSEMs) offer researchers a way to investigate how the interplay of individual behaviours makes aggregation possible and leads to different aggregate level behaviours. In this thesis I argue that in order to model species-specific aggregate behaviours, modellers must pay attention to three properties of animals that uniquely determine their aggregating behaviour: the perceptual mechanisms of the animal, the movement abilities of the animal and the cognitive mechanisms that connect perception to action. The existing modelling framework, the Attraction Repulsion (AR) framework, does not allow for an adequate representation of these three facets of individual aggregating behaviour. As an alternative I present a new modelling framework, the Perception Cognition Action (PCA) framework. I use this framework to create an experimentally based, experimentally validated computer model of Danio rerio aggregate behaviour. I further demonstrate, more generally, the importance of modelling perception by analysing the behaviour of a theoretical aggregate model that varies the perceptual abilities of agents and determines the effect that this has on aggregate cohesion. Lastly, I explore the interplay between perception, cognition and movement by varying individual perception, cognition and action properties. I analyse the behaviour of the resulting aggregate with a focus on how the interaction between perception, cognition and action determines aggregate properties.