Navigating to the Best Policy in Markov Decision Processes.
Aymen Al MarjaniAurélien GarivierAlexandre ProutièrePublished in: CoRR (2021)
Keyphrases
- markov decision processes
- optimal policy
- policy iteration
- markov decision process
- average reward
- state and action spaces
- infinite horizon
- partially observable
- average cost
- reinforcement learning
- finite state
- action space
- state space
- reward function
- finite horizon
- dynamic programming
- total reward
- decision processes
- discount factor
- policy iteration algorithm
- discounted reward
- decision problems
- transition matrices
- markov decision problems
- state dependent
- reachability analysis
- expected reward
- decision theoretic planning
- factored mdps
- long run
- partially observable markov decision processes
- reinforcement learning algorithms
- semi markov decision processes
- multistage
- policy evaluation
- control policies
- planning under uncertainty
- continuous state spaces
- stationary policies
- model based reinforcement learning
- sufficient conditions
- actor critic
- risk sensitive