A note on optimality conditions for continuous-time Markov decision processes with average cost criterion.
Xianping GuoKe LiuPublished in: IEEE Trans. Autom. Control. (2001)
Keyphrases
- average cost
- optimality conditions
- markov decision processes
- state space
- stationary policies
- optimal control
- nonlinear programming
- finite state
- optimal policy
- markov chain
- finite horizon
- policy iteration
- linear programming
- dynamic programming
- reinforcement learning
- approximate dynamic programming
- infinite horizon
- lower level
- average reward
- long run
- initial state
- risk sensitive
- optimality criterion
- markov decision process
- dynamical systems
- partially observable
- action space
- sample size
- sufficient conditions
- finite number
- action sets
- linear program
- variational inequalities
- markov decision problems
- higher level
- sensitivity analysis
- multistage
- reward function