Pseudometrics for State Aggregation in Average Reward Markov Decision Processes.
Ronald OrtnerPublished in: ALT (2007)
Keyphrases
- markov decision processes
- average reward
- state space
- discounted reward
- optimal policy
- total reward
- policy iteration
- state action
- markov decision process
- reinforcement learning
- finite state
- discount factor
- action space
- stochastic games
- partially observable
- semi markov decision processes
- dynamic programming
- reinforcement learning algorithms
- long run
- state abstraction
- optimality criterion
- average cost
- planning under uncertainty
- real time dynamic programming
- infinite horizon
- reward function
- factored mdps
- model free
- decision theoretic planning
- decision processes
- state and action spaces
- state variables
- least squares
- initial state
- heuristic search
- markov chain