A Lyapunov-based Approach to Safe Reinforcement Learning.
Yinlam ChowOfir NachumEdgar A. Duéñez-GuzmánMohammad GhavamzadehPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- adaptive control
- function approximation
- stability analysis
- class of nonlinear systems
- reinforcement learning algorithms
- optimal control
- markov decision processes
- state space
- sufficient conditions
- data sets
- model free
- learning process
- nonlinear systems
- learning algorithm
- machine learning
- reinforcement learning methods
- robotic control
- dynamic programming
- function approximators
- temporal difference learning
- transition model
- neural network