Modeling of Adaptive Cyber Physical Systems using Aspect-oriented Approach

This paper proposes an aspect-oriented approach to modeling adaptive cyber physical system (CPS) using Petri nets. The core concerns of CPSs are described as device model and task model, and dynamic variations of system behaviors or environment conditions are extracted as crosscutting concerns. The models of runtime inspection as well as device adaptation and task adaptation are designed as aspects nets. For the device adaptation strategy, fault types are analyzed and the control loop concept is integrated to form the adaptation aspect model. For the task adaptation, a rescheduling method using PSO-Pareto algorithm to find the best solution of the backup devices is proposed. Via well-defined rules, these aspect nets can be weaved with the core concern nets into a comprehensive adaptive CPS model. By theoretical analysis and a case study, we show the modeling approach is feasible and flexible, which simplifies the design of adaptive CPSs.