Towards evolution-based autonomy in large-scale systems

Anderson, Damien and Harvey, Paul and Kaneta, Yusaku and Papadopoulos, Petros and Rodgers, Philip and Roper, Marc; (2022) Towards evolution-based autonomy in large-scale systems. In: GECCO 2022 Companion - Proceedings of the 2022 Genetic and Evolutionary Computation Conference. GECCO 2022 Companion - Proceedings of the 2022 Genetic and Evolutionary Computation Conference . ACM, USA, 1924–1925. ISBN 9781450392686 (https://doi.org/10.1145/3520304.3533975)

[thumbnail of Anderson-etal-GECCO-2022-Towards-evolution-based-autonomy-in-large-scale-systems]
Preview
Text. Filename: Anderson_etal_GECCO_2022_Towards_evolution_based_autonomy_in_large_scale_systems.pdf
Accepted Author Manuscript
License: Strathprints license 1.0

Download (259kB)| Preview

Abstract

To achieve truly autonomous behaviour systems need to adapt to not only previously unseen circumstances but also previously unimagined ones. A hierarchical evolutionary system is proposed which is capable of displaying the emergent behaviour necessary to achieve this goal. This paper reports on the practical challenges encountered in implementing this proposed approach in a large-scale open-source system. In "Evolutionary Autonomous Networks"[3], Harvey et al. present their vision for an autonomous network utilising online evolution to achieve guided emergent behaviour. Their approach presents a radical strategy to improve a network's ability to cope with uncertainty and change. The aim of this paper is to present the initial steps we have taken towards realising this vision by examining how evolutionary-based autonomy may be introduced into an example large-scale system-a content delivery network (CDN). This goal is not without its challenges as there are significant hurdles imposed by the design and implementation of the CDN which must be overcome.