The Capacity of Mixed and One-Sided Gaussian Interference Channels

TL;DR

This paper fully characterizes the capacity regions of one-sided and degraded Gaussian interference channels, introduces new bounds for mixed interference regimes, and confirms the optimality of partial interference decoding depending on system parameters.

Contribution

It provides a complete capacity characterization for one-sided and degraded Gaussian interference channels and introduces new outer bounds for mixed regimes, advancing understanding of interference management.

Findings

Capacity region of one-sided Gaussian interference channel fully characterized.

New outer bounds for mixed interference regimes established.

Decoding part of interference is optimal depending on system parameters.

Abstract

This paper adds to the understanding of the capacity region of the Gaussian interference channel. To this end, the capacity region of the one-sided Gaussian interference channel is first fully characterized. This is accomplished by introducing a new representation of the Han-Kobayashi region and providing a new outer bound on the capacity region of this channel which is tight both in the weak and strong interference regimes. In light of this capacity result, the capacity region of the degraded Gaussian interference channel is established immediately. Next, by combining the capacity regions of the one-sided interference channels in the weak and strong interference regimes, new outer bounds on the capacity region of the interference channel are introduced, in various interference regimes. It is then proved that the outer bound corresponding to the mixed interference regime, in which one of the receivers is subject to strong interference while the other one suffers from weak interference, is tight in a broad range of this regime. The new capacity results, on the whole, confirm the optimality of decoding part of the interference and treating the rest as noise. The optimum amount to be decoded varies from 0 to 100% of the interfering signal, depending on the relative importance of the users' rates (i.e., the ratio of weights in the weighted sum-rate), their transmission powers, and the gain of the weak interference link. Optimal values are found explicitly, based on the above parameters.