Efficient simulation tools for digital twinning in cyber-physical systems:new challenges and a proposed road map

Creating digital twins (DT) of cyber-physical systems (CPS) presents unique challenges that current modelling and simulation (M&S) tools are not fully equipped to address. While some aspects of this challenge have been well explored, such as multi-domain modelling, real-time data synchronisation, and predictive analytics using AI and machine learning, others have received less attention. Two important, yet in our opinion underexplored, areas are (i) the handling of large systems of differential and algebraic equations (DAE) coupled with event-based dynamics, as Finite State Machine (FSM), and (ii) the need for not only efficient simulation code but also for efficiency in the process of generating that code. We discuss the challenges related to these topics in a CPS-based context, and sketch out a roadmap for developing solutions that enhance the effectiveness and scalability of M&S tools.