CAREER: Approximations and Bounds for Stationary and Transient Characteristics of a Class of Queueing Networks

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

The objective of this CAREER research is to develop analytical models of queuing networks that are prevalent in manufacturing systems. Specifically, develop approximations and bounds for certain characteristics of these queuing networks. Some key performance measures to be considered are mean, variance, tail probability of transient and stationary sojourn times and covariance of transient and stationary waiting times. Unfortunately, it is very hard to get closed-form expressions for such performance measures unless the network has a very simple structure and it is operating under very restrictive assumptions. To this end, expansion formulae will be provided that can be used to numerically approximate these characteristics. The analytical models developed during this project can be eventually used to design new job release and scheduling strategies that will lead to improvements in manufacturing operations and ultimately increase the competitiveness of U.S. manufacturers. This CAREER teaching plan aims to improve engineering education by focusing on new course development projects, development of case studies and other course materials and involving undergraduate and master's students in research. In particular, the educational effort will focus on the development of course materials for the restructured undergraduate stochastic processes course and the design of a new graduate level course that will focus on the stochastic models in manufacturing. In addition, it will concentrate on enriching the students' educational experience by devising active learning techniques and involving undergraduate and masters students in research.