A graph-based developmental swarm representation and algorithm

Abstract

Modelling natural processes requires the implementation of an expressive representation of the involved entities and their interactions. We present swarm graph grammars (SGGs) as a bio-inspired modelling framework that integrates aspects of formal grammars, graph-based representation and multi-agent simulation. In SGGs, the substitution of subgraphs that represent locally defined agent interactions drive the computational process of the simulation. The generative character of formal grammars is translated into an agent's metabolic interactions, i.e. creating or removing agents from the system. Utilizing graphs to describe interactions and relationships between pairs or sets of agents offers an easily accessible way of modelling biological phenomena. Property graphs emerge through the application of local interaction rules; we use these graphs to capture various aspects of the interaction dynamics at any given step of a simulation. © 2010 Springer-Verlag Berlin Heidelberg.