Asymptotic properties of constrained Markov Decision Processes.
Eitan AltmanPublished in: ZOR Methods Model. Oper. Res. (1993)
Keyphrases
- markov decision processes
- asymptotic properties
- finite state
- policy iteration
- state space
- fixed point
- optimal policy
- reinforcement learning
- dynamic programming
- transition matrices
- planning under uncertainty
- partially observable
- risk sensitive
- decision processes
- average cost
- reinforcement learning algorithms
- action sets
- state and action spaces
- infinite horizon
- finite horizon
- action space
- decision theoretic planning
- factored mdps
- reachability analysis
- semi markov decision processes
- average reward
- model based reinforcement learning
- reward function
- collaborative filtering