---
title: "Biologically-Inspired Self-Assembly of Two-Dimensional Shapes Using Global-to-Local Compilation"
authors:
  - "Attila Kondacs"
doi: ""
url: "https://ijcai.org/proceedings/2003/93"
pdf: "https://ijcai.org/proceedings/2003/0093.pdf"
venue: "Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence"
year: 2003
pages: "633-637"
track: ""
keywords:
---

# Biologically-Inspired Self-Assembly of Two-Dimensional Shapes Using Global-to-Local Compilation

Attila Kondacs

*Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence*, pages 633-637.

## Abstract

In this ·paper, we present a programming language approach for the assembly of arbitrary twodimensional shapes by decentralized, identicallyprogrammed agents. Our system compiles a predetermined global shape into a program that instructs these agents to grow the shape via replication and location-based control mechanisms. In the globalto-local compilation phase, an input shape is decomposed into a network of covering-discs. The disc network parameterizes the agent program, a biologically-inspired framework allowing agents to amorphously produce the shape using replication and local interaction. Our system is robust to random agent failure, and regenerates in the event of region death.