Continuous-time Markov decision processes under the risk-sensitive average cost criterion.
Qingda WeiXian ChenPublished in: Oper. Res. Lett. (2016)
Keyphrases
- risk sensitive
- average cost
- markov decision processes
- state space
- optimal control
- markov decision chains
- stationary policies
- finite state
- finite horizon
- optimal policy
- reinforcement learning
- infinite horizon
- markov chain
- dynamic programming
- policy iteration
- decision processes
- planning under uncertainty
- average reward
- initial state
- optimality criterion
- partially observable
- action space
- control policy
- markov decision problems
- long run
- reinforcement learning algorithms
- belief state
- markov decision process
- reward function
- finite number
- multistage
- lower bound