Reinforcement Learning for Partially Observable Dynamic Processes: Adaptive Dynamic Programming Using Measured Output Data

By Frank L. Lewis, Kyriakos G. Vamvoudakis. IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), 2011

497 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Linear Quadratic Tracking Control of Partially-Unknown Continuous-Time Systems Using Reinforcement Learning (2014, methodology): “For linear systems which are the focus of this technical note, PI has been effectively used to solve the optimal regulator problem for both DT systems [11]–[13] and CT systems [14]–[16].”
Computational adaptive optimal control for continuous-time linear systems with completely unknown dynamics (2012, background): “…been extended and applied to many different areas, such as nonzero-sum games (Al-Tamimi et al., 2017), networked control systems (Xu et al., 2018), and optimal output feedback control designs (Lewis & Vamvoudakis, 2017), as well as multi-objective optimal control problems (Wei, Zhang, & Dai, 2009).”
Model-Free Optimal Tracking Control via Critic-Only Q-Learning (2016, background): “Note that most of the existing related works [12], [24], [42], [51], [66], [68], [79], [80] belong to the on-policy learning framework or use on-policy implementation.”
Reinforcement Learning and Feedback Control: Using Natural Decision Methods to Design Optimal Adaptive Controllers (2012, background): “Design of an ADP controller that uses only output feedback is given in [50].”
Reinforcement-Learning-Based Robust Controller Design for Continuous-Time Uncertain Nonlinear Systems Subject to Input Constraints (2015, methodology): “Consequently, RL is often applied to adaptive optimal controller designs [24]–[37].”