Safe Exploration in Markov Decision Processes .
Teodor Mihai MoldovanPieter AbbeelPublished in: ICML (2012)
Keyphrases
- markov decision processes
- model based reinforcement learning
- interval estimation
- optimal policy
- state space
- finite state
- dynamic programming
- reachability analysis
- reinforcement learning
- transition matrices
- policy iteration
- average cost
- decision theoretic planning
- factored mdps
- infinite horizon
- risk sensitive
- action sets
- reinforcement learning algorithms
- planning under uncertainty
- average reward
- action space
- finite horizon
- state abstraction
- decision processes
- markov decision process
- reward function
- semi markov decision processes
- partially observable
- state and action spaces