Robust topological policy iteration for infinite horizon bounded Markov Decision Processes.
Willy Arthur Silva ReisLeliane Nunes de BarrosKarina Valdivia DelgadoPublished in: Int. J. Approx. Reason. (2019)
Keyphrases
- policy iteration
- markov decision processes
- infinite horizon
- optimal policy
- finite horizon
- state space
- dynamic programming
- finite state
- markov decision process
- partially observable
- average cost
- reinforcement learning
- policy evaluation
- sample path
- reinforcement learning algorithms
- optimal control
- average reward
- single item
- planning under uncertainty
- markov decision problems
- transition matrices
- long run
- model free
- factored mdps
- stochastic games
- decision processes
- partially observable markov decision processes
- actor critic
- dec pomdps
- discounted reward
- policy iteration algorithm