Analysis of Accuracy of Queuing Models


Ping Xiao

Oral Defence Date: 



TH 434


Profs. Dujmovic and Wong


The primary goal of this project is to develop a simulator of four types of queuing systems: single server model, multiple servers model, feedback queuing model, and interactive system model with general distribution of interarrival time and general service time distributions. In the case of exponential distributions the work of each simulator can be validated using analytic models. In cases that use general distribution our goal is to evaluate the robustness of exponential queuing models as predictors of performance indicators of such systems. The simulators produce numerical results and histograms for various distributions such as interarrival time, inter-departure time, response time, and so on. This technique is used to demonstrate how feedback in queuing networks creates non-exponential flows of service requests even in the case of Poisson arrival process. The results obtained using simulation are compared with analytic models and used for systematic analysis of errors and a study of robustness of queuing models of open and closed queuing networks.


Simulation, single server, multiple servers model, feedback queuing model, interactive system, robustness of queuing models


Ping Xiao