Conservation Equations and Their Application to Queuing Simulations

Abstract

In this paper we continue our study of efficient estimators for simulated queuing systems. We previously showed that it is more efficient to estimate the first moments of all the usual measures of performance from an estimate of mean wait in queue than to estimate them directly. Here we derive some new conservation equations that are then used to obtain efficient estimators of the variances of the usual measures of performance. We also introduce an estimator that is more efficient than the usual estimator for the probability that an arriving customer is blocked.