Parameter estimation in continuous and discrete time queues using queue length and waiting time data

Parameter estimation is investigated for single-server queueing systems based on limited information, specifically, customer waiting times and queue lengths.; Maximum likelihood and conditional least squares estimators are developed for the discrete-time Geo/D/1 model using waiting time data. Strong consistency and asymptotic normality of these estimators are derived. Also using waiting time data, the maximum likelihood estimator for the discrete-time Geo/Geo/1 model is developed. The existence, and strong consistency of the estimator are obtained.; Using queue length data, maximum likelihood, conditional least squares and maximum quasilikelihood estimators are derived for Geo/ D/1, Geo/Geo/1 and continuous-time M/E_k/1 queueing models. Their strong consistency and asymptotic normality are established.; Simulations are performed to study the finite sample behavior of all the estimators developed. Estimators based on full data, queue length data and waiting time data are compared.