ELEN E6970 Resource Allocation and Networking Games

Course Benefits

Professor Lazar

Applicable Degree Programs

Most courses 4000-level and above can be credited to all degree programs. All courses are subject to advisor approval.
Lecturer Professor Aurel A. Lazar
Class location: 1127 S.W. Mudd
Office hours: Mondays and Wednesdays, 11:00-12:00 AM EST
Office phone: +1 212 854 1747
Email address: aurel "at" ee.columbia.edu
Class Web Site: Offered by the COMET Group
Mailing list:
Day and time: Mondays and Wednesdays, 9:35-10:50 AM EST
Credits: 4.5 points
Prerequisites: A graduate level course in computer communication networks (e.g., EE E6761) or a course in stochastic processes (e.g., EE E6711), or the instructor's approval. No prior knowledge of game theory is required; game-theoretic fundamentals will be covered in class.
Description: Future multimedia networks will support a wide variety of service classes, each with its own traffic characteristics and quality of service requirements. Efficient allocation of network resources under quality of service guarantees is one of the key factors in such networking environments. Allocation of resources is determined by the interaction among various controllers (e.g., flow controllers, routers) that are distributed within the network. This interaction can be modeled as game. The goal of the course is to present the state-of-the-art in resource allocation with emphasis on networking games.

Resource allocation in single-class networks: flow control, routing. Multi-class networks: classification of network control algorithms, motivation of the game-theoretic formulation, the greedy algorithm; Network control games: flow control, routing, virtual path bandwidth reservation; Applications to network design and management: motivation, the Braess paradox, architecting network resources and user flows; Network pricing: pricing of congestible network resources, economics of the Internet, pricing of services, allocation of the wireless spectrum. A programming project using Java, Matlab or Mathematica is required.

Required text(s): none
Reference text(s): Drew Fudenberg and Jean Tirole, Game Theory, The MIT Press, Cambridge, MA., 1992.
Roger B. Myerson, Game Theory: Analysis of Conflict, Harvard University Press, Cambridge, MA, 1991.
Martin J. Osborne and Ariel Rubinstein, A Course in Game Theory,, The MIT Press, Cambridge, MA, 1994.

Updated class notes will be available under the URL http://comet.ctr.columbia.edu/courses/elen_e6970/2002/notes.html

Homework(s): Four homework assignments - all optional. The homeworks will help you to prepare for the final exam. Submitting homework assigments earns you extra credit.
Projects:

Students are expected to complete a major programming project during the course of the semester. A detailed project plan has to be submitted before the midterm break. The plan can not be more then four pages in length including figures and references. It has to be submitted in pdf format.

Paper(s): See "Project(s)"
Midterm exam: None
Final exam: 24 hours take home
Grading: Choice between project 2/3 and final 1/3 or project 1/3 and final 2/3. You have to make a decision about your choice before the midterm break. The project tests creativity whereas the final tests analytical ability.
Hardware requirements: Access to a PC or workstation. Solaris, Windows and Apple Power PC OS are strongly preferred.
Software requirements: Access to a Java platform, Matlab or Mathematica.
Homework submission: Usually, by electronic mail.