Partially observable queueing systems with controlled service rates under a discounted optimality criterion.
Yofre H. GarcíaSaúl Díaz-InfanteJ. Adolfo Minjárez-SosaPublished in: Kybernetika (2021)
Keyphrases
- queueing systems
- partially observable
- service rates
- optimality criterion
- average reward
- markov decision processes
- arrival rate
- heavy traffic
- infinite horizon
- long run
- queue length
- optimal policy
- asymptotically optimal
- reinforcement learning
- queueing networks
- state dependent
- state space
- decision problems
- service times
- steady state
- reward function
- dynamic programming
- policy iteration
- queueing model
- finite state
- stationary distribution
- single server
- average cost
- call center
- dynamical systems
- markov decision process
- model free
- multistage
- partially observable markov decision processes
- evaluation function
- lead time
- sufficient conditions
- markov chain
- markov random field
- belief state
- multi agent