Rare-Event Simulation of Non-Markovian Queueing Networks Using a State-Dependent Change of Measure Determined Using Cross-Entropy.
Pieter-Tjerk de BoerPublished in: Ann. Oper. Res. (2005)
Keyphrases
- queueing networks
- state dependent
- cross entropy
- product form
- rare events
- steady state
- stationary distribution
- single server
- heavy traffic
- importance sampling
- maximum likelihood
- log likelihood
- arrival rate
- markov chain
- queue length
- scoring function
- long run
- evaluation metrics
- reinforcement learning
- similarity measure
- ranking functions
- optimal policy
- probabilistic model
- search space