Integrating developmental and conventional Markov decision processes: An application to robotic navigation.
Shuqing ZengYanhua ChenPublished in: IJCNN (2013)
Keyphrases
- markov decision processes
- robotic navigation
- finite state
- optimal policy
- reinforcement learning
- state space
- dynamic programming
- transition matrices
- planning under uncertainty
- decision theoretic planning
- reinforcement learning algorithms
- reachability analysis
- decision processes
- policy iteration
- factored mdps
- finite horizon
- action space
- model based reinforcement learning
- partially observable
- average cost
- average reward
- total reward
- action sets
- infinite horizon
- risk sensitive
- reward function
- markov decision process
- expected reward
- state abstraction
- continuous state spaces
- policy evaluation
- decision problems