Markov Decision Processes with Average-Value-at-Risk criteria.
Nicole BäuerleJonathan OttPublished in: Math. Methods Oper. Res. (2011)
Keyphrases
- markov decision processes
- risk sensitive
- average cost
- state space
- discounted reward
- finite state
- optimal policy
- dynamic programming
- reinforcement learning
- decision processes
- transition matrices
- finite horizon
- reinforcement learning algorithms
- reachability analysis
- optimality criterion
- decision theoretic planning
- planning under uncertainty
- infinite horizon
- partially observable
- average reward
- stationary policies
- optimal control
- factored mdps
- markov decision process
- initial state
- action space
- state and action spaces
- model based reinforcement learning
- state abstraction
- reward function
- action sets
- finite number