Provably Safe Reinforcement Learning with Step-wise Violation Constraints.
Nuoya XiongYihan DuLongbo huangPublished in: CoRR (2023)
Keyphrases
- step wise
- reinforcement learning
- learning algorithm
- constraint satisfaction
- function approximation
- constraint violations
- state space
- supervised learning
- markov decision processes
- constrained optimization
- resource constraints
- model free
- machine learning
- worst case
- reinforcement learning algorithms
- theoretical guarantees