Achieving Target State-Action Frequencies in Multichain Average-Reward Markov Decision Processes.
Dmitry KrassO. J. VriezePublished in: Math. Oper. Res. (2002)
Keyphrases
- average reward
- state action
- markov decision processes
- optimal policy
- policy iteration
- stochastic games
- semi markov decision processes
- long run
- reinforcement learning
- finite state
- state space
- reinforcement learning algorithms
- dynamic programming
- total reward
- markov decision process
- average cost
- infinite horizon
- partially observable
- discounted reward
- state and action spaces
- model free
- action space
- decision problems
- markov chain
- reward function
- random walk
- objective function