Disciplined Exploitation of Emergent Properties

Digital systems are becoming increasingly complex, requiring significantly more effort and resources in order to be designed, implemented, and maintained. In the last decade, industry and academia alike share the concern that in the near future engineers will have to face unprecedented levels of complexity. Similarly, the belief that traditional engineering approaches will be insufficient for coping with systems of such complexity is gaining increasingly more supporters. An alternative approach suggests the use of implicit engineering techniques which could lead to complex global-level behaviours by focusing solely on the local, or individual, level. The quality of rising macroscopic behaviours which are irreducible, or non-trivial to reduce, to any microscopic properties is more widely known as emergence, especially in the fields of complex and multi-agent systems.

This work aims to investigate the possibility of engineering systems which harness, in intentional and disciplined ways, beneficial emergent properties. An experimental framework is being proposed to assist system designers towards that goal. This framework is based on the results and experience gained by Paunovski during the design of the Emergent Distributed Bio-Organisation (EDBO) case study. EDBO has demonstrated a number of beneficial emergent properties rising out of simple, bio-inspired, local interactions. The original implementation of the EDBO case study is closely coupled with a custom simulation platform; both developed by the same author. This work provides a basis for separating the EDBO case study from this combined implementation, by documenting it concisely and defining it in a formal manner. This formal model allows for rigorous testing and enables other authors to reuse the EDBO principles in their systems. The model is validated informally through animation and it serves as the basis of an independent implementation which cross-validated many of EDBO's original findings.