Principal Investigators: Bernd Meyer and Chris Reid

Partner Investigators: Tim Landgraf and Iain Couzin

Effective decision making and coordination in all forms of large, complex groups requires well-organised and structured communication: information must be communicated and processed selectively where it is relevant, in functional subgroups of the collective. Yet, at some point this information must be integrated across the whole group to achieve a group-wide balance of demands and available resources. Our understanding of how communication network structures enable large groups to achieve this balance between local and global information processing is limited at best.

This project investigates when and how structured and modular communication networks emerge in large, complex societies to effectively balance local and global information processing. We use task allocation in social insect colonies as an established model system for our investigation.

Task allocation happens in a completely decentralised fashion without central coordination and control. In fact, “task allocation” is a misnomer: tasks are not centrally allocated to individuals, but individuals make independent decisions about task selection based on the information locally available to them. Task selection closely relates to modular, localised information processing. Foraging decisions are made by one subgroup, brood-care decisions by another, and so on. Clearly, information processing for task allocation cannot be totally compartmentalised. Managing simultaneous tasks requires integration at the colony level to achieve a global balance of demands and available resources. The enormous ecological success of social insects is partly attributed to effective task allocation and thus to such integration of information processing. However, its mechanisms are poorly understood. Modular communication is highly likely to be a core factor in this integration, because it can strike a balance between global and local processing.

The project uses cutting-edge computational techniques, in particular AI-based computer vision, to empirically analyse the structure of communication networks in social insect colonies and how these vary with with colony size and task complexity. We hope to explain the functional significance of the observed variations by building network-based mathematical models and computer simulations of colony-level decision making. We also aim to identify evolutionary pathways for the emergence of these networks using evolutionary game theory.

This work is funded by the Australian Research Council and conducted at Monash University and Macquarie University in collaboration with the Biorobotics Lab at Free University Berlin and the Department of Collective Behaviour at the Max Planck Institute of Animal Behaviour and the University of Konstanz.