Towards a generic observer/controller architecture for Organic Computing.

TLDR

Technical scenarios such as automotive or production systems increasingly involve many components in dynamic networks, making explicit design of all possible behaviours infeasible, and Organic Computing envisions self‑organising systems that adapt robustly to changing environments. The paper aims to design a generic system architecture that permits self‑organisation while providing sufficient control over emergent global behaviour. The architecture enables self‑organisation and incorporates mechanisms to react to unexpected emergent global behaviour in these technical systems.

Abstract

Technical scenarios in areas like automotive or production systems will increasingly consist of a large number of components cooperating in potentially unlimited and dynamically changing networks to satisfy the functional requirements of their execution environment. Due to the high complexity it will be impossible to explicitly design the behaviour of the components for every potentially arising situation. Therefore, it will be necessary to leave an adequate degree of freedom allowing for a self-organised behaviour. Organic Computing (OC) has developed the vision of selforganising systems adapting robustly to dynamically changing environments without running out of control. This paper focuses on the design of a generic system architecture which allows for self-organisation but at the same time enables adequate reactions to control the – sometimes completely unexpected – emerging global behaviour of these self-organised technical systems.