Abstract
Applications of queueing models for analyzing real-life scenarios are based on the fundamental premise that parameter values are known. This is a rarity and more often than not, such parameters require to be estimated. Any analyst with field level experience will corroborate this. How are such estimates to be obtained? What are the procedures to be followed? How stable are those? These and a few other related questions of fundamental practical interest are dealt with in this paper. The queue we intend to deal with is the M/M/1 queueing model. In other words this paper deals with statistical inference with regard to the classical single server Markovian queueing model. Inference is discussed from Bayesian perspective as it has certain advantages. Point estimates of performance measures are obtained. Predictive distributions of number of customers in system are derived along with their moments. Interval estimates and tests of hypotheses are also presented.

This publication has 11 references indexed in Scilit: