End-to-End Reinforcement Learning of Koopman Models for eNMPC of an Air Separation Unit

A new study highlights the effectiveness of a reinforcement learning method for training Koopman surrogate models, which can significantly enhance the performance of economic nonlinear model predictive control in air separation units. This advancement is crucial as it demonstrates the method's scalability to more complex scenarios, paving the way for improved efficiency and optimization in industrial applications. The implications of this research could lead to better resource management and cost savings in various sectors.