A STUDY OF THE OUTPUT PROCESS OF A QUEUEING SYSTEM AND ITS COVARIANCE EFFECT IN TANDEM QUEUES.
Item
-
Title
-
A STUDY OF THE OUTPUT PROCESS OF A QUEUEING SYSTEM AND ITS COVARIANCE EFFECT IN TANDEM QUEUES.
-
Identifier
-
AAI8801692
-
identifier
-
8801692
-
Creator
-
CHIOU, SU-CHAO.
-
Contributor
-
Giorghois P. Sphicas
-
Date
-
1987
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Business Administration, Management | Computer Science
-
Abstract
-
A wide range of Erlang phases is considered on the interarrival time distribution and the processing time distribution in the Erlangian queues (E{dollar}\sb{lcub}\rm j{rcub}{dollar} /E{dollar}\sb{lcub}\rm k{rcub}{dollar} /1) to represent a more flexible queueing system. The primary focus of this study is on the departure process of such system and its covariance effect on the next station in tandem queues.;This study uses computer simulation as the primary methodology incorporating a set of numerical procedures. The interdeparture intervals of the consecutive departing transactions are first collected for the purpose of analysis. These are fitted into a hypothesized Gamma distribution with estimated scale parameter and shape parameter. These estimated parameters are then tested by Chi-square and Kolmogorov-Smirnov tests of the goodness-of-fit testing procedures. The test statistics are found to be very statistically significant.;These estimated parameters of the output process are then fitted in a power function of the input and processing time parameters with strong empirical evidence. The characteristics of lag-1 covariance and correlation structure of the departure process are also obtained and fitted into a quadratic function of the input and processing time parameters with very significant empirical support.;The covariances of the consecutive interdeparture times are found to be non-zero. They are positive on some cases, and are negative on some other cases. The t-test is then employed to test the significance level and the sign of the correlation of interdeparture intervals. The degree of significance and the sign of the correlation is found to be mainly determined by the system utilization and Erlang phases. The effect of these significant positive and negative covariances on the waiting time measurement in the next station of a tandem queue is then examined. It is found that the positive lag-1 covariance results in an underestimate on the waiting time measurement; and negative lag-1 covariance results in an overestimate.;The results of this kind of study are useful in the design and control of a serial production system. With the knowledge of the underestimate or overestimate on the performance measurement, the control process of the serial production system can be simplified.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Buisness
