Reachable Space Characterization of Markov Decision Processes with Time Variability.
Junhong XuKai YinLantao LiuPublished in: Robotics: Science and Systems (2019)
Keyphrases
- markov decision processes
- action space
- finite state
- optimal policy
- transition matrices
- dynamic programming
- state space
- decision theoretic planning
- policy iteration
- reinforcement learning
- reachability analysis
- finite horizon
- average cost
- planning under uncertainty
- infinite horizon
- partially observable
- action sets
- reward function
- average reward
- reinforcement learning algorithms
- semi markov decision processes
- state and action spaces
- search space
- total reward
- probabilistic planning
- decision diagrams
- risk sensitive
- factored mdps
- model based reinforcement learning
- markov decision process