Hierarchical Bayesian Inference for Time-Varying M|M|1 Queuing Systems
摘要
Hierarchical Bayes estimations and predictive inferences for a M|M|1 queueing system are carried out, assuming that the traffic intensities vary over time. Two different models are considered for the cases when the time points have no pattern and when time points follow a pattern during the day. Posterior distributions of traffic intensities and predictive distributions of future queue sizes are derived. Bayes estimators of varying traffic intensities along with posterior inferences using Gibbs sampler are obtained using variational Bayesian method. Two simulated examples are considered to illustrate the numerical computations.