Trading Performance for Stability in Markov Decision Processes.
Tomás BrázdilKrishnendu ChatterjeeVojtech ForejtAntonín KuceraPublished in: LICS (2013)
Keyphrases
- markov decision processes
- finite state
- transition matrices
- optimal policy
- state space
- policy iteration
- dynamic programming
- reinforcement learning
- reachability analysis
- decision theoretic planning
- finite horizon
- reinforcement learning algorithms
- action sets
- average cost
- partially observable
- infinite horizon
- risk sensitive
- markov decision process
- planning under uncertainty
- model based reinforcement learning
- action space
- decision processes
- average reward
- monte carlo
- factored mdps
- state and action spaces