In this note, we present a critical analysis of machine learning techniques for applications involving optimal (feedback) control. Specifically, we will focus on the question of using reinforcement learning and other similar techniques in providing provably stable optimal controllers.
展开▼
