Nonhomogeneous Markov Decision Processes with Borel State Space-The Average Criterion with Nonuniformly Bounded Rewards.
Xianping GuoJianyong LiuKe LiuPublished in: Math. Oper. Res. (2000)
Keyphrases
- markov decision processes
- state space
- average cost
- discounted reward
- stationary policies
- reinforcement learning
- optimal policy
- finite state
- dynamic programming
- reinforcement learning algorithms
- transition matrices
- reward function
- heuristic search
- average reward
- decision theoretic planning
- markov chain
- policy iteration
- factored mdps
- reachability analysis
- planning problems
- markov decision process
- model based reinforcement learning
- planning under uncertainty
- optimality criterion
- sequential decision making under uncertainty
- decision processes
- finite horizon
- dynamical systems
- state and action spaces
- partially observable
- action space
- fully observable
- state variables
- markov decision problems
- macro actions
- action sets
- total reward
- partially observable markov decision processes
- infinite horizon