Design and Control of Service Centers

Lee, Nelson

Download PDF

Request Version for Screen Reader

Last Modified

March 19, 2019

Creator

Lee, Nelson
- Affiliation: College of Arts and Sciences, Department of Statistics and Operations Research

Abstract

A service center is a facility with multiple heterogeneous servers providing specialized service to multiple types of customers. Design and control problems of service centers arise in many practical applications such as cloud computing, data centers, health care facilities, call centers, etc. With the motivation of reducing energy consumption in data centers, this dissertation investigates the design and control problems from three different perspectives that are applicable to the service centers in general. The first study provides decision models to determine optimal static assignment and routing policies, explicitly taking into account the stochastic fluctuations of demand along with the autocorrelations and cross-correlations of the different traffic streams. We consider several possible performance measures and formulate the design problem as a mixed integer nonlinear programming problem. We also develop an efficient heuristic algorithm to enhance scalability. We observe numerically that the optimal routing policy tries to combine the negatively correlated traffic streams, and separate the positively correlated traffic streams. The second study is motivated by the virtual computing labs (VCL) at UNC-Chapel Hill and North Carolina State University that provide remote access to users allowing them to reserve and use a virtual computer with their desired software application set. This study proposes an energy efficient scheduling and operating policy for such computing facilities. It is of general applicability in large scale computing systems. Using Erlang-B model, we provide a server scheduling and operating policy that results in solutions that satisfy the given performance requirements with minimum energy consumption. We also evaluate our policy with actual data collected from the VCL at North Carolina State University. Finally, the third study considers the problem of optimal control of a multi-server queue with controllable arrival and service rates. The cost structure includes customer holding cost which is a non-decreasing convex function of the number of customers in the system, server operating cost which is a non-decreasing convex function of the chosen service rate, and system operating reward which is a non-decreasing concave function of the chosen arrival rate. We formulate the problem as a continuous-time Markov decision process, and derive structural properties of the optimal control policies under both discounted cost and average cost criterions.

Date of publication