Cyber-physical systems (CPS), e.g. applied in the areas of IT-based traffic control, communication networks and logistics systems, are generally subject to a steady change in the ambient conditions, the objectives, but also the internal conditions of a CPS. E.g. the control of road traffic and the management of communication networks to achieve optimized system behaviour must respond to errors, changed user requirements, or attacks.
In order to achieve an optimized system behaviour in spite of this dynamic setting, CPS are equipped with a flexible structure. Both the existing components including their parameterization and networking can be changed in order to realize the desired system behaviour. Thus, such flexible CPSs are subject to a continuous re-configuration task to optimize their overall behaviour.
The most successful techniques for configuring and reconfiguring technical systems are currently based on knowledge-based methods. In this approach, the technical possibilities of a system and domain knowledge are formally described in a representation language, and optimized configurations are created from the description by using complete and correct reasoning systems. The advantages of this approach are an increase in the solution quality while at the same time achieving a considerable reduction of the development and maintenance costs of (re)configurators compared to procedural programming.
Today, the technology of knowledge-based (re)configuration with its advantages is not available for CPS. The goal of DynaCon is to advance the currently available techniques of knowledge-based (re)configuration in a twofold manner, such that this technology is enabled for a broad range of CPS:
First, dynamic reconfiguration in a time-sensitive manner will be achieved. The input, i.e. the description of the current configuration, the configuration actions as well as the system objectives, is not regarded as static, but as a data stream. The adaptation of the configuration to the environmental conditions is carried out continuously, where unnecessary adaptations should be avoided.
Second, the efficiency of the currently available reasoning procedures will be significantly increased so that large instances of (re)configuration problems can be solved under tight time constraints; this poses a major challenge.
Within DynaCon, the improvements are demonstrated by automated (re)configuration in four selected application areas:
DynaCon enables a continuous re-configuration of IT-based road traffic control systems; this results in better management of urban road traffic.
DynaCon allows a rapid re-configuration of communication networks to control power generation; this increases availability.
DynaCon permits automatic and rapid response to anomalies, overloads and failures in communication networks of commercial network service providers; this increases productivity and service quality.
DynaCon enables a rapid response to disturbances in the operation of railway logistics; this reduces transport costs.