Competition and Cooperation in Multi-user Communication Environments

Mr. Wei Yu

Electrical Engineering Department
Stanford University, CA

Monday, April 29th, 11:00 AM, ENS 637

weiyu@stanford.edu


Abstract

A communication environment with multiple transmitters and multiple receivers is inherently a competitive environment. The aim of this talk is to illustrate the role of competition and the value of cooperation in multi-user communication environments from an information theoretical perspective. Various scenarios will be treated, including the multiple access channel where receivers cooperate, the broadcast channel where transmitters cooperate, and the interference channel where neither transmitters nor receivers cooperate.

There are three main results in this talk. First, we show that in a Gaussian multiple access channel with multiple transmit and receive antennas, the optimum transmission strategy that maximizes the sum capacity can be found by an iterative water-filling procedure, where each user competitively maximizes its own rate. Thus, a competitive optimum in a Gaussian multiple access channel is also a global optimum. Second, we look at the long-standing non-degraded broadcast channel problem, and show that in a multi-antenna Gaussian broadcast channel, the sum capacity is achieved at a saddle-point of a mutual information game, where the transmitter chooses a transmit strategy to maximize the mutual information, and "Nature" chooses a fictitious noise correlation to minimize the mutual information. Thus, the sum capacity of a Gaussian vector broadcast channel corresponds to a competitive equilibrium. Third, we show that in a Gaussian interference channel, although a competitive optimum is not necessarily the global optimum, it leads to a desirable operating point.

Implications for practical wireless and wireline system design will be discussed. Several novel multi-user transmission and network optimization techniques will be presented.

Biography

Wei Yu received his B.A.Sc in Computer Engineering and Mathematics from the University of Waterloo, Canada in 1997. He will receive a Ph.D. degree in electrical engineering from Stanford University in 2002. His research interests include network information theory, communication theory, and wireless and wireline broadband access networks.


A list of Telecommunications and Signal Processing Seminars is available at from the ECE department Web pages under "Seminars". The Web address for the Telecommunications and Signal Processing Seminars is http://signal.ece.utexas.edu/seminars