On the Minimum Pair Approach for Average Cost Markov Decision Processes with Countable Discrete Action Spaces and Strictly Unbounded Costs.
Huizhen YuPublished in: SIAM J. Control. Optim. (2020)
Keyphrases
- average cost
- markov decision processes
- action space
- continuous state
- continuous state spaces
- state action
- state space
- state and action spaces
- finite state
- optimal policy
- dynamic programming
- finite horizon
- reinforcement learning
- markov decision process
- infinite horizon
- policy iteration
- control policies
- markov decision problems
- risk sensitive
- average reward
- planning under uncertainty
- action sets
- decision processes
- reinforcement learning algorithms
- initial state
- partially observable
- control policy
- reward function
- decision making
- machine learning
- function approximators
- total cost
- linear program