Agent-Based Modeling and Simulation of Hybrid Cyber-Physical Systems

Cyber-physical production systems are becoming more complex and heterogeneous in the frame of Internet of Cyber-Physical Things and Industry 4.0 environments. Behavior of such systems is difficult to analyze and control due to the system scale and emergency aspects. Multi-agent paradigm is a suitable formalism for modeling these kinds of systems and simulation modeling is important for getting an insight into the system, for synthesizing its control, and for analyzing performance and efficiency of systems under various circumstances. This paper is focused on simulation modeling for complex hybrid cyber-physical production systems. They combine continuous-time parts and discrete-event parts and the goal is to bring a methodology for addressing simulation for such complex hybrid systems. The paper presents two possible approaches how to unify access to both paradigms for system modeling and it explains how to combine them meaningfully. The proposed solution is based on the application of the traditional time-driven approach on the level of individual agents simulating cyber-physical components or sub-systems, and the dual value-driven approach frequently denoted as Quantized State Systems method in such a way that the benefits of both approaches are gained. The proposed approach simplifies synchronization and improves stability of coupled simulations by self-optimization of sampling periods for synchronizing the simulation agents for cyber-physical components.