Transfer and Reinforcement Learning Based Production Control

Constantly increasing complexity and growing information densities in production systems open up potentials for the application of machine learning methods. So-called reinforcement learning is particularly suitable for implementing autonomous agentbased control systems. However, the application of this becomes more difficult in changing production systems. It is shown for the first time that the transfer learning approach is useful for production control systems with reinforcement learning. © 2022 Walter de Gruyter GmbH, Berlin/Boston.