An Information-Spectrum Approach to the Capacity Region of General Interference Channel

TL;DR

This paper introduces a general formula for the capacity region of interference channels using an information-spectrum approach, providing insights into achievable rates and potential improvements over simple noise-treating strategies.

Contribution

It presents a novel capacity region formula for general interference channels based on spectral inf-mutual information rates, extending beyond traditional methods.

Findings

The capacity region is the union of rectangles determined by spectral inf-mutual information rates.

Treating interference as noise can be improved by considering interference structure.

Numerical results for Gaussian channels with convolutional codes support the theoretical insights.

Abstract

This paper is concerned with general interference channels characterized by a sequence of transition (conditional) probabilities. We present a general formula for the capacity region of the interference channel with two pairs of users. The formula shows that the capacity region is the union of a family of rectangles, where each rectangle is determined by a pair of spectral inf-mutual information rates. Although the presented formula is usually difficult to compute, it provides us useful insights into the interference channels. For example, the formula suggests us that the simplest inner bounds (obtained by treating the interference as noise) could be improved by taking into account the structure of the interference processes. This is verified numerically by computing the mutual information rates for Gaussian interference channels with embedded convolutional codes.