Anytime Guarantees for Reachability in Uncountable Markov Decision Processes.
Kush GroverJan KretínskýTobias MeggendorferMaximilian WeiningerPublished in: CONCUR (2022)
Keyphrases
- markov decision processes
- state space
- optimal policy
- dynamic programming
- reinforcement learning
- finite state
- initial state
- reinforcement learning algorithms
- policy iteration
- transition matrices
- partially observable
- heuristic search
- finite horizon
- decision theoretic planning
- factored mdps
- reachability analysis
- average reward
- average cost
- planning under uncertainty
- action sets
- risk sensitive
- decision processes
- state and action spaces
- machine learning
- model based reinforcement learning
- continuous state spaces
- markov decision problems
- learning algorithm
- markov decision process
- least squares