Accession Number : AD0704788

Title :   MAXIMUM-LIKELIHOOD ESTIMATES FOR QUEUES WITH STATE-DEPENDENT SERVICE,

Corporate Author : GEORGE WASHINGTON UNIV WASHINGTON D C PROGRAM IN LOGISTICS

Personal Author(s) : Goyal,Tarsaim L. ; Harris,Carl M.

Report Date : 24 MAR 1970

Pagination or Media Count : 44

Abstract : Methods are indicated for the maximum-likelihood estimation of the arrival and service parameters of ergodic queueing processes with service times whose distributions are state dependent. Two basic methods of sampling are considered. For the first, only the queue size at each departure point is observed, while for the second, the time between successive departures and the lengths of the idle periods are noted along with the queue size. Based upon each sampling method, procedures are developed for deriving estimates both when the effect of initial queue size on the likelihood function is ignored and when it cannot be. Three specific models of this semi-Markov process are presented, along with a numerical example illustrating some of the results, based on data generated by simulating a known model. (Author)

Descriptors :   (*QUEUEING THEORY, MATHEMATICAL PREDICTION), STATISTICAL DISTRIBUTIONS, SAMPLING, MATHEMATICAL MODELS, STOCHASTIC PROCESSES, NUMERICAL ANALYSIS

Subject Categories : Operations Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE