Thompson Sampling for Parameterized Markov Decision Processes with Uninformative Actions.
Michael GimelfarbMichael Jong KimPublished in: CoRR (2023)
Keyphrases
- markov decision processes
- decision theoretic planning
- partially observable
- state and action spaces
- action space
- action sets
- decision processes
- reward function
- macro actions
- finite state
- optimal policy
- reinforcement learning
- state space
- policy iteration
- transition matrices
- finite horizon
- dynamic programming
- reachability analysis
- average reward
- infinite horizon
- model based reinforcement learning
- average cost
- reinforcement learning algorithms
- markov decision process
- state transitions
- stochastic games
- planning under uncertainty
- risk sensitive
- multiagent reinforcement learning
- least squares
- learning algorithm