Information Theory of Wireless Networks: A Deterministic Approach

Dr. David Tse
University of California, Berkeley

Wednesday, November 7, 2007

Gaussian channels are commonly used models for multiuser wireless communication. Unfortunately, the capacity of multiuser Gaussian channels is unknown even for simple networks such as the single-relay channel and the two-user interference channel. To make further progress in understanding how to optimally communicate over these and more general networks, we propose a deterministic channel model which focuses on the interaction between the users rather than the noise in the system. We show:

  1. the analytical simplicity of this model by computing the capacities of several interference and relay networks based on this deterministic model
  2. how the insights from the deterministic model can be translated into finding near-optimal strategies for the Gaussian counterpart.