Modelling morphogenesis as an amorphous computation

Abstract

This thesis presents a programming-language viewpoint for morphogenesis, the process of shape formation during embryological development. We model morphogenesis as a self-organizing, self-repairing amorphous computation and describe how we can program large-scale shape formation by giving local instructions to cell-like objects. Our goal is to simulate systems that display properties, like robustness, regeneration, and evolvability, that are present in biological systems but ordinarily not present in computer systems. Consistent with the theory of facilitated variation from evolutionary biology, we find that many of these properties can be introduced and conserved by a hierarchical organization of growth specification.