The use of variance reduction, relative error and bias in testing the performance of M/G/1 retrial queues estimators in Monte Carlo simulation.
Kenza TamitiMegdouda Ourbih-TariAbdelouhab AlouiKhelidja IdjisPublished in: Monte Carlo Methods Appl. (2018)
Keyphrases
- variance reduction
- monte carlo simulation
- relative error
- single server
- monte carlo
- markov chain
- retrial queue
- importance sampling
- steady state
- queue length
- confidence intervals
- gradient estimation
- queueing networks
- bias variance decomposition
- state dependent
- special case
- service times
- original data
- stationary distribution
- queueing model
- sample size
- parallel machines
- heavy traffic
- state space
- arrival rate
- markov chain monte carlo
- machine learning
- test set
- least squares
- feature space
- search algorithm