EE 599 Webpage --- Stochastic Network Optimization


Instructor:
Michael Neely (mjneely AT usc.edu, 213-740-3505, EEB 520)

Syllabus:
A syllabus with a more detailed course description is found here: EE 599 Course Syllabus

Brief Course Description:
This course presents a theory of stochastic optimization and cross-layer control for modern data networks with time varying channels, user mobility, and randomly arriving traffic. Emphasis will be on wireless networks, with applications also to computer networks and mixed wireless/wireline systems. We model networks as queueing systems with general transmission rate capabilities determined by the physical properties of each network element. Explicitly including queues in the model not only provides a more complete cross-layer perspective, but facilitates the design of opportunistic resource allocation, routing, and flow control decisions. Mathematical techniques of Lyapunov Drift, Lyapunov Optimization, and Virtual Cost Queues will be introduced and used as important tools in the design of optimal dynamic control algorithms. This theory unifies notions of static convex programming and stochastic network control, and complements existing courses on linear and convex programming for networks as well as existing courses on data networks and queueing theory.

Prerequisites:
EE 465 or EE 562a. Familiarity with queueing systems and Markov chains is required. Advanced networking courses such as EE 549 and/or EE 550 are strongly recommended pre- or co- requisites, but not required.

Course Topics:
Network Layer Capacity, Optimal Control of Wireless and Ad-Hoc Mobile Networks, Opportunistic Resource Allocation, Routing, Flow Control, Minimum Energy Networking, Pricing, General Utilities and Constraints, Queue Stability, Energy-Delay and Utility-Delay Tradeoffs, Complexity Issues and Imperfect Scheduling.

Textbook:
The course will use a combination of instructor handouts, lecture notes, journal articles, and the following text:

L. Georgiadis, M. J. Neely, L. Tassiulas, "Resource Allocation and Cross-Layer Control in Wireless Networks," Foundations and Trends in Networking, Vol. 1, no. 1, pp. 1-144, 2006.

A printed and bound version of this article is available at a 45% discount from Now Publishers.

This can be obtained by entering the promotional code NET001001 on the order form at now publishers. You will then pay only $49.50 including postage.


Course Handouts:
  1. This is a set of background math notes on Limsups, Sets, and Convexity.
  2. This is a list of papers on course topics: Paper List.
  3. This is Problem Set 1: PS1.
  4. This is a summary of results for the first 3 lectures: Summary of Results.
  5. This is a sample C program that implements a B/B/1 queue: BB1.c.
  6. This is a link to the Stochastic Network Optimization Page, which has a few selected papers in the field together with a brief description of each.


Back to Michael Neely Home Page