A note on deterministic approximation of discounted Markov decision processes.
Hugo Cruz-SuárezEvgueni GordienkoRaúl Montes-de-OcaPublished in: Appl. Math. Lett. (2009)
Keyphrases
- markov decision processes
- stationary policies
- optimal policy
- dynamic programming
- policy iteration
- infinite horizon
- state space
- finite state
- policy iteration algorithm
- reinforcement learning
- average cost
- finite horizon
- decision theoretic planning
- average reward
- transition matrices
- reinforcement learning algorithms
- planning under uncertainty
- risk sensitive
- reachability analysis
- markov decision process
- action space
- policy evaluation
- action sets
- linear program
- model based reinforcement learning
- decision processes
- discount factor
- partially observable
- decision diagrams
- decision problems
- factored mdps
- state and action spaces
- semi markov decision processes
- least squares
- collaborative filtering
- efficient computation
- approximation methods
- partially observable markov decision processes
- queueing networks