Reversible Markov Decision Processes with an Average-Reward Criterion.
Randy CogillCheng PengPublished in: SIAM J. Control. Optim. (2013)
Keyphrases
- average reward
- markov decision processes
- optimality criterion
- optimal policy
- policy iteration
- markov chain
- semi markov decision processes
- stochastic games
- state space
- discounted reward
- finite state
- decision theoretic planning
- long run
- dynamic programming
- reinforcement learning
- state and action spaces
- total reward
- hierarchical reinforcement learning
- partially observable
- reinforcement learning algorithms
- factored mdps
- decision processes
- markov decision process
- action space
- infinite horizon
- state abstraction
- random walk
- discount factor
- decision problems
- model free
- optimal solution
- policy gradient
- planning under uncertainty
- average cost