Limiting discounted-cost control of partially observable stochastic systems.
Onksimo Hernandez-LermaRosario RomeraPublished in: CDC (2000)
Keyphrases
- partially observable
- stochastic systems
- infinite horizon
- markov decision processes
- average cost
- optimal control
- state space
- dynamic programming
- reinforcement learning
- decision problems
- control system
- finite horizon
- dynamical systems
- long run
- sample path
- stochastic models
- control method
- policy iteration
- finite state
- sufficient conditions
- total cost
- control strategy
- expected cost
- optimal policy
- initial state
- chaotic systems
- special case