Horizon-Free Reinforcement Learning for Latent Markov Decision Processes.
Runlong ZhouRuosong WangSimon S. DuPublished in: CoRR (2022)
Keyphrases
- markov decision processes
- reinforcement learning
- optimal policy
- reinforcement learning algorithms
- state space
- state and action spaces
- discount factor
- finite state
- policy iteration
- model based reinforcement learning
- action space
- dynamic programming
- function approximation
- reachability analysis
- state abstraction
- transition matrices
- planning under uncertainty
- action sets
- decision processes
- markov decision process
- finite horizon
- partially observable
- average cost
- decision theoretic planning
- factored mdps
- latent variables
- learning algorithm
- control problems
- model free
- infinite horizon
- multi agent
- machine learning
- average reward
- approximate dynamic programming
- continuous state spaces
- optimal control