Safe Reinforcement Learning in Constrained Markov Decision Processes.
Akifumi WachiYanan SuiPublished in: ICML (2020)
Keyphrases
- markov decision processes
- reinforcement learning
- reinforcement learning algorithms
- optimal policy
- state space
- finite state
- state and action spaces
- policy iteration
- dynamic programming
- function approximation
- partially observable
- action space
- decision processes
- planning under uncertainty
- markov decision process
- average cost
- transition matrices
- finite horizon
- reachability analysis
- learning algorithm
- average reward
- control problems
- state abstraction
- factored mdps
- action sets
- infinite horizon
- stochastic games
- reward function
- decision theoretic planning
- model based reinforcement learning
- optimal control
- policy evaluation
- real time dynamic programming
- policy iteration algorithm
- discounted reward
- multi agent
- decentralized control
- continuous state spaces
- continuous state