Stabilizing Dynamical Systems via Policy Gradient Methods.
Juan C. PerdomoJack UmenbergerMax SimchowitzPublished in: NeurIPS (2021)
Keyphrases
- dynamical systems
- policy gradient methods
- natural actor critic
- reinforcement learning methods
- policy gradient
- nonlinear systems
- actor critic
- robot arm
- dynamic systems
- partially observable markov decision processes
- fixed point
- control law
- state space
- reinforcement learning problems
- partially observable
- nonlinear dynamical systems
- neural network
- real valued
- predictive state representations
- dynamic programming
- machine learning