Finite state approximations of Markov decision processes with general state and action spaces.
Naci SaldiTamás LinderSerdar YükselPublished in: ACC (2015)
Keyphrases
- markov decision processes
- finite state
- state and action spaces
- action space
- optimal policy
- state space
- reinforcement learning
- dynamic programming
- policy iteration
- average reward
- average cost
- policy evaluation
- markov decision process
- action sets
- reinforcement learning algorithms
- decision processes
- infinite horizon
- partially observable
- planning under uncertainty
- partially observable markov decision processes
- reward function
- markov chain
- stationary policies
- dynamical systems
- machine learning