Nonparametric estimation of the service time distribution in discrete-time queueing networks (Q2186641)

The paper under review provides nonparametric estimation of the service time distributions at the nodes of the discrete queueing network, in which the nodes are \(GI/G/\infty\) queueing systems, in the case that only external movements of customers are observable. The paper extends the existing approaches based on covariance function by providing functional central limit theorem for the resultant cross-covariance estimators. For those cross-covariance estimators, a simple procedure for ensuring tightness in the function space of absolutely summable sequences is established under quite general settings. The accuracy of the estimator is illustrated by its comparison with the simulation results.