On the existence of stationary optimal policies for the average cost control problem of linear systems with abstract state-feedback.
Alessandro N. VargasJoão Bosco Ribeiro do ValPublished in: CDC (2008)
Keyphrases
- average cost
- optimal policy
- stationary policies
- linear systems
- control policy
- state space
- markov decision processes
- sufficient conditions
- optimal control
- long run
- finite state
- infinite horizon
- initial state
- markov decision process
- finite number
- control policies
- state dependent
- finite horizon
- decision problems
- dynamic programming
- dynamical systems
- multistage
- linear program
- markov decision chains
- reinforcement learning
- total cost
- inventory models
- setup cost
- policy iteration
- state variables
- action space
- holding cost
- linear programming
- markov chain
- control system
- partially observable
- pid controller
- average reward
- lost sales