LSTM-empowered reinforcement learning in Bi-level optimal control for nonlinear systems with uncertain dynamics.

This paper introduces a bi-level optimization framework for the optimal control of nonlinear continuous-time systems with uncertain dynamics, seamlessly integrating Long Short-Term Memory (LSTM) networks with an actor-critic reinforcement learning (R…

Preview only. Read the full abstract at the source