Approximation of average cost optimal policies for general Markov decision processes with unbounded costs.
Evgueni GordienkoRaúl Montes-de-OcaJ. Adolfo Minjárez-SosaPublished in: Math. Methods Oper. Res. (1997)
Keyphrases
- average cost
- markov decision processes
- optimal policy
- finite state
- finite horizon
- long run
- state space
- infinite horizon
- reinforcement learning
- policy iteration
- special case
- dynamic programming
- average reward
- risk sensitive
- decision problems
- setup cost
- finite number
- total cost
- stationary policies
- optimal control
- holding cost
- multistage
- initial state
- control policies
- policy evaluation
- decision processes
- inventory models
- action sets
- control policy
- reinforcement learning algorithms
- partially observable
- linear programming
- state dependent
- markov decision process
- data mining
- search algorithm
- semi markov decision processes
- queueing networks
- function approximation
- linear program
- sufficient conditions
- expected reward
- discount factor