Markov decision processes with risk-sensitive criteria: dynamic programming operators and discounted stochastic games.
Rolando Cavazos-CadenaEmmanuel Fernández-GaucherandPublished in: CDC (2001)
Keyphrases
- markov decision processes
- risk sensitive
- stochastic games
- dynamic programming
- average reward
- optimality criterion
- optimal policy
- state space
- infinite horizon
- finite state
- optimal control
- average cost
- reinforcement learning algorithms
- policy iteration
- finite horizon
- linear programming
- markov decision problems
- decision processes
- partially observable
- reinforcement learning
- markov decision process
- machine learning
- multistage
- action space