Optimization, Game and Layering in Communication Networks

Course Description

This course discusses various equilibrium solution concepts and convergent algorithms in optimization and game theory, and their applications to network design and control. The underlying theme is “network protocols as distributed algorithms achieving various equilibria”. The objective is to introduce mathematically rigorous tools for analyzing current network protocols and designing new ones. Topics will include: Equilibrium solution concepts and convergent algorithms in optimization and game theory, the utility maximization framework of TCP congestion control, layering as optimization decomposition, path algebra and routing, contention control, power control, and distributed mechanism design for network problems.

Units and Grading

6 units (2-0-4); pass/fail, or letter grade.

Wednesday 4 PM - 6 PM; 214 Steele

Lijun Chen (chen@cds.caltech.edu)
Office hours: by appointment
Office: room 5 Steele; Ext. 3367

Announcements

• 03 Oct: Syllabus is posted.
• 03 Oct: The organizing meeting will be on Wednesday, October 03, 2007 in 214 Steele at 4:00 PM. The lectures will begin in the second week.

Course Material

• Overview
• Lecture 1: Static Games and Classical Mechanism Design (Lecture Notes)
  Readings: Drew Fudenberg and Jean Tirole, Game Theory, MIT Press, 1991, pp.3--44.
  Classic Mechanism Design, D. C. Parkes, Chapter 2 in PhD thesis, Iterative Combinatorial Auctions: Achieving Economic and Computational Efficiency, University of Pennsylvania, 2001.
  Mechanism Theory, M. Jackson, Encyclopedia of Life Support Systems, 2003.

Useful Links

• Dr. Maryam Fazel's course on Convex Optimization and Applications.