Industrial Benchmark Bridging the Gap Between Artificial and Real-world Problem Attribute Datasets

Industrial Benchmark is a dataset that bridges the gap between free, documented, and active artificial problems and the properties of real industrial problems. The dataset is about the realities of offline and online RL and is used to find the RL algorithms that are best suited for real-world applications. The dataset aims to provide interpretable RL training scenarios and detailed understanding of the learning process of the methods. (The resulting Industrial Data (IB) has been publicly made available to the RL community by publishing its Java and Python code on Github, including an OpenAI Gym wrapper)