Fundamental Limits of Communications in Interference Networks-Part IV: Networks with a Sequence of Less-Noisy Receivers

TL;DR

This paper characterizes the sum-rate capacity of interference networks with a sequence of less-noisy receivers using a novel sequential outer bound technique, extending understanding from two-receiver to multi-receiver scenarios.

Contribution

It introduces a new sequential Csiszar-Korner identity technique to establish outer bounds and fully characterizes sum-rate capacity for large multi-receiver interference networks with less-noisy conditions.

Findings

Sum-rate capacity characterized for large multi-receiver networks.

Novel sequential outer bound technique developed.

Outer bounds are proven to be sum-rate optimal under certain conditions.

Abstract

In this fourth part of our multi-part papers, classes of interference networks with a sequence of less-noisy receivers are identified for which a successive decoding scheme achieve the sum-rate capacity. First, the two-receiver networks are analyzed: it is demonstrated that the unified outer bounds derived in Part III of our multi-part papers are sum-rate optimal for network scenarios which satisfy certain less-noisy conditions. Then, the multi-receiver networks are considered. These networks are far less understood. One of the main difficulties in the analysis of such scenarios is how to establish useful capacity outer bounds. In this paper, a novel technique requiring a sequential application of the Csiszar-Korner identity is developed to establish powerful single-letter outer bounds on the sum-rate capacity of multi-receiver interference networks which satisfy certain less-noisy conditions. By using these outer bounds, a full characterization of the sum-rate capacity is derived for general interference networks of arbitrary large sizes with a sequence of less-noisy receivers. Some generalizations of these outer bounds are also presented each of which is efficient to obtain the exact sum-rate capacity for various scenarios.