Online-Learning-Based Economic MPC of Switched Nonlinear Systems*

Abstract

This work develops a Lyapunov-based economic model predictive control (LEMPC) scheme that utilizes recurrent neural networks (RNNs) with online update to optimize the economic benefits of switched nonlinear systems subject to a prescribed switching schedule. We first develop an initial offline-learning RNN using historical operational data, and then update RNN models using real-time data to improve model prediction accuracy. The generalized error bound for RNNs updated online with non-independent and identically distributed (non-i.i.d.) data samples is first derived. Subsequently, by incorporating the online update of RNNs within LEMPC, probabilistic closed-loop stability and economic optimality are achieved simultaneously for switched nonlinear systems accounting for the RNN generalized error bound. A chemical process example with scheduled mode transitions is used to demonstrate that the closed-loop economic performance under LEMPC can be improved using online learning of RNNs.