Swarm Grammars: Modeling Computational Development through Highly Dynamic Complex Processes

Abstract

We have developed swarm grammars (SGs) as an integrated representation of artificial swarms and developmental models. SGs have evolved in three steps: basic swarm grammars \citeMammen:2006ve,Jacob:2007qf,Mammen:2007ly, extended swarm grammars \citeMammen:2008fk,Mammen:2008bh and swarm graph grammars (SGGs). In this thesis, we present each of these types of SGs together with their respective motivations. We dedicate chapters to the evolutionary exploration of SGs and their application in interdisciplinary works in the fields of art \citeMammen:2008uq and architecture \citeMammen:2008uq,Mammen:2008fk. The breeding methods we applied for evolving architectural design required the introduction of complexity measures in SGs. These in turn triggered a study on swarm networks \citeMammen:2008vn that promoted the development of SGGs. Through SGGs, swarm grammars have matured into a universal, bio-inspired modeling framework for complex developmental systems.