Individual based and mean-field modelling of direct aggregation

TL;DR

This paper introduces novel individual-based and mean-field models of biological aggregation driven solely by density-dependent stochasticity reduction, with mathematical analysis and numerical simulations demonstrating pattern formation and steady states.

Contribution

It presents new models where aggregation emerges without explicit forces, based on stochasticity reduction, and provides rigorous analysis and simulations of these models.

Findings

Existence of weak solutions for the first-order model

Conditions for pattern formation identified

Numerical simulations confirm theoretical results

Abstract

We introduce two models of biological aggregation, based on randomly moving particles with individual stochasticity depending on the perceived average population density in their neighbourhood. In the first-order model the location of each individual is subject to a density-dependent random walk, while in the second-order model the density-dependent random walk acts on the velocity variable, together with a density-dependent damping term. The main novelty of our models is that we do not assume any explicit aggregative force acting on the individuals; instead, aggregation is obtained exclusively by reducing the individual stochasticity in response to higher perceived density. We formally derive the corresponding mean-field limits, leading to nonlocal degenerate diffusions. Then, we carry out the mathematical analysis of the first-order model, in particular, we prove the existence of weak solutions and show that it allows for measure-valued steady states. We also perform linear stability analysis and identify conditions for pattern formation. Moreover, we discuss the role of the nonlocality for well-posedness of the first-order model. Finally, we present results of numerical simulations for both the first- and second-order model on the individual-based and continuum levels of description.