Actor-Based Designs for Distributed Self-organisation Programming

Self-organisation and collective adaptation are highly desired features for several kinds of large-scale distributed systems including robotic swarms, computational ecosystems, wearable collectives, and Internet-of-Things systems. These kinds of distributed processes, addressing functional and non-functional aspects of complex socio-technical systems, can emerge in an engineered/controlled way from (re)active decentralised activity and interaction across all physical and logical system devices. In this work, we study how the Actors programming model can be adopted to support collective self-organising behaviours. Specifically, we analyse the features of the Actors model, such as reactivity, asynchrony, and locality, that are instrumental for implementing the adaptive coordination of large-scale systems, and discuss potential actor-based designs, from simple ad-hoc implementation of algorithms to a full-fledged general toolkit. In particular, the approach is incarnated in the aggregate computing paradigm, which stands as a comprehensive engineering approach for self-organisation. This is based on Akka, and can be fully programmed in the Scala programming language thanks to the ScaFi aggregate computing toolkit.