Digital twins (DT) are the virtual counterpart of a physical system. Successfully deployed applications
of DT’s are available in today’s cutting-edge technology but limited to specific industries owing to the fact that
development of a DT is interminable depending on the system’s level of complexity. Discrete event specification
(DEVS) is one of dynamic system modeling formalism that can be used to model a wide variety of dynamic
systems of interest. Reinforcement learning (RL) is a machine learning technique that focuses on how to react
to specific states of an environment—that is, how to map situations and motions of an object into actions—–in
order to maximize a reward signal. The proposed research will develop an industry oriented transposable
DT framework via integrating reinforcement learning algorithms with DEVS modeling and simulation. The
framework will be applicable to a DT of interest of an industrial system. The results of the research will be
used to evaluate the potency of the framework in improving the operation, and maintenance of the industrial
system thereby contributing an innovative way of using DT in today’s internet of things equipped industrial
systems.
