Effective Wireless Scheduling via Hypergraph Sketches

TL;DR

This paper introduces a novel framework for approximating wireless network capacity under the physical interference model, significantly improving bounds by simplifying complex interference relationships into conflict graphs.

Contribution

It presents the first approximation algorithms for wireless capacity in the physical model, using hypergraph sketches to simplify interference analysis with provably optimal bounds.

Findings

Approximation bounds are at most doubly logarithmic in link length and rate diversity.

Complex interference can be simplified into conflict graphs with minimal accuracy loss.

The obtained approximation is proven to be optimal for any conflict graph approach.

Abstract

An overarching issue in resource management of wireless networks is assessing their capacity: How much communication can be achieved in a network, utilizing all the tools available: power control, scheduling, routing, channel assignment and rate adjustment? We propose the first framework for approximation algorithms in the physical model of wireless interference that addresses these questions in full. The approximations obtained are at most doubly logarithmic in the link length and rate diversity. Where previous bounds are known, this gives an exponential improvement (or better). A key contribution is showing that the complex interference relationship of the physical model can be simplified, at a small cost, into a novel type of amenable conflict graphs. We also show that the approximation obtained is provably the best possible for any conflict graph formulation.