The effect of estimation methods on performance measures for queuing systems (M/M/1)

Abstract

In this research, we review one of the probability distributions related to queuing theory (Poisson distribution, Exponential distribution) and the estimation methods for its parameters, which are the traditional estimation methods (Maximum Likelihood, White). The basic aspect of this research is the experimental aspect, as the experimental aspect includes employing the simulation method in a way Monte Carlo is used to generate data with an exponential distribution, then find estimators for its measurement parameter and determine the preference of these estimators that relate to the distribution of service times in queuing systems (M/M/1) using one of the statistical measures, which is the measure of the mean square error (MSE). ) In order to demonstrate the superiority of these estimators with respect to the two estimation methods and sample sizes and several default values for the measurement parameter, then creating estimators for the performance measures for the queuing system (M/M/1). The researchers concluded that there is an effect of the estimation methods on the performance measures, as the method (White) is the best for performance metrics.