Abstract

Aggregation is one of the most widespread phenomena in animal groups and often represents a collective dynamic response to environmental conditions. In social species the underlying mechanisms mostly obey self-organized principles. This phenomenon constitutes a powerful model to decouple purely social components from ecological factors. Here we used a model of cockroach aggregation to address the problems of sensitivity of collective patterns and control of aggregation dynamics. The individual behavioural rules and the emergent collective patterns were previously quantified and modelled by Jeanson et al.. We first present the diverse spatio-temporal patterns of a derived model in response to parameter changes, either involving social or non-social interactions. This sensitivity analysis is then extended to evaluate the evolution of these patterns in mixed societies of sub-populations with different behavioural parameters. Simple linear or highly non-linear collective responses emerge. We discuss their potential application to control animal populations by infiltration of biomimetic autonomous robots that mimic cockroach behaviour. We suggest that detailed behavioural models are a prerequisite to do so.