Simulation-based policy generation using large-scale Markov decision processes.
Christopher W. ZobelWilliam T. SchererPublished in: IEEE Trans. Syst. Man Cybern. Part A (2001)
Keyphrases
- markov decision processes
- optimal policy
- policy iteration
- markov decision process
- average reward
- finite horizon
- infinite horizon
- action space
- partially observable
- average cost
- state space
- finite state
- reward function
- decision processes
- dynamic programming
- reinforcement learning
- state and action spaces
- transition matrices
- policy iteration algorithm
- expected reward
- policy evaluation
- reachability analysis
- planning under uncertainty
- decision problems
- discounted reward
- markov decision problems
- reinforcement learning algorithms
- factored mdps
- decision theoretic planning
- sufficient conditions
- long run
- continuous state spaces
- risk sensitive
- stationary policies
- control policies
- total reward
- multistage
- state dependent
- semi markov decision processes
- model based reinforcement learning
- partially observable markov decision processes
- action sets