The vanishing discount approach in Markov chains with risk-sensitive criteria.
Rolando Cavazos-CadenaEmmanuel Fernández-GaucherandPublished in: IEEE Trans. Autom. Control. (2000)
Keyphrases
- markov chain
- risk sensitive
- optimality criterion
- optimal control
- steady state
- markov decision processes
- state space
- finite state
- transition probabilities
- utility function
- average reward
- random walk
- monte carlo
- markov decision problems
- stationary distribution
- markov model
- markov processes
- model free
- transition matrix
- control policies
- probabilistic automata
- average cost
- decision theoretic
- expected utility
- evaluation function
- dynamic programming
- reinforcement learning
- optimal policy
- machine learning