Approximate Fixed Point Iteration with an Application to Infinite Horizon Markov Decision Processes.
Anthony AlmudevarPublished in: SIAM J. Control. Optim. (2008)
Keyphrases
- fixed point
- infinite horizon
- markov decision processes
- optimal policy
- policy iteration
- finite horizon
- policy evaluation
- state space
- finite state
- partially observable
- sufficient conditions
- dynamic programming
- markov decision process
- average cost
- reinforcement learning
- single item
- dynamical systems
- reinforcement learning algorithms
- belief propagation
- planning under uncertainty
- average reward
- long run
- markov decision problems
- dec pomdps
- multistage
- inventory level
- linear programming
- search algorithm
- bayesian networks
- learning algorithm
- sample path
- machine learning