Reinforcement Learning for Statistical Process Control in Manufacturing

The main concept of the authors is to place Reinforcement Learning (RL) into various fields of manufacturing. As one of the first implementations, RL for Statistical Process Control (SPC) in production is introduced in the paper; it is a promising approach owing to its adaptability and the continuous ability to perform. The widely used Q-Table method was applied for get more stable, predictable, and easy to overview results. Therefore, quantization of the values of the time series to stripes inside the control chart was introduced. Detailed elements of the production environment simulation are described and its interaction with the reinforcement learning agent are detailed. Beyond the working concept for adapting RL into SPC in manufacturing, some novel RL extensions are also described, like the epsilon self-control of exploration–exploitation ratio, Reusing Window (RW) and the Measurement Window (MW). In the production related transformation, the main aim of the agent is to optimize the production cost while keeping the ratio of good products on a high level as well. Finally, industrial testing and validation is described that proved the applicability of the proposed concept.