Optimal policies of M(t)/M/c/c queues with two different levels of servers.
Ali TirdadWinfried K. GrassmannJavad TavakoliPublished in: Eur. J. Oper. Res. (2016)
Keyphrases
- optimal policy
- service rates
- state dependent
- markov decision processes
- asymptotically optimal
- decision problems
- dynamic programming
- reinforcement learning
- finite horizon
- infinite horizon
- finite state
- state space
- multistage
- long run
- dynamic programming algorithms
- initial state
- average reward reinforcement learning
- sufficient conditions
- service times
- markov decision process
- average reward
- average cost
- partially observable markov decision processes
- control policies
- policy iteration
- serial inventory systems
- stationary distribution
- partially observable
- arrival rate
- queueing networks