SMPL: Simulated Industrial Manufacturing and Process Control Learning Environments
This work addresses a gap for researchers in industrial manufacturing and process control by providing tools for benchmarking, but it is incremental as it builds on existing dynamics models without introducing new control methods.
The authors tackled the lack of high-fidelity simulations and standard APIs for applying deep reinforcement learning to industrial manufacturing control by developing a library with five simulation environments based on published dynamics models, and they benchmarked various reinforcement learning algorithms to facilitate future research.
Traditional biological and pharmaceutical manufacturing plants are controlled by human workers or pre-defined thresholds. Modernized factories have advanced process control algorithms such as model predictive control (MPC). However, there is little exploration of applying deep reinforcement learning to control manufacturing plants. One of the reasons is the lack of high fidelity simulations and standard APIs for benchmarking. To bridge this gap, we develop an easy-to-use library that includes five high-fidelity simulation environments: BeerFMTEnv, ReactorEnv, AtropineEnv, PenSimEnv and mAbEnv, which cover a wide range of manufacturing processes. We build these environments on published dynamics models. Furthermore, we benchmark online and offline, model-based and model-free reinforcement learning algorithms for comparisons of follow-up research.