Variance minimization for constrained discounted continuous-time MDPs with exponentially distributed stopping times.
Jun FeiEugene A. FeinbergPublished in: Ann. Oper. Res. (2013)
Keyphrases
- markov decision processes
- state space
- finite horizon
- optimal policy
- infinite horizon
- average cost
- average reward
- optimal control
- markov decision process
- finite state
- stationary policies
- dynamic programming
- reinforcement learning
- policy iteration
- markov chain
- objective function
- action space
- reinforcement learning algorithms
- markov processes
- factored mdps
- state and action spaces
- partially observable
- semi markov decision processes
- correlation coefficient
- covariance matrix
- dynamical systems
- standard deviation
- planning problems
- planning under uncertainty
- decision theoretic planning
- real time dynamic programming
- discount factor
- machine learning
- discounted reward
- finite number
- learning algorithm