A Greedy link scheduler for Wireless Networks having Gaussian Broadcast and Multiple Access Channels

TL;DR

This paper introduces a greedy link scheduling algorithm for multihop wireless networks with broadcast and multiple access capabilities, improving throughput by leveraging multiuser channels and extending interference models.

Contribution

It proposes a novel Multiuser Greedy Maximum Weight algorithm and develops local pooling conditions for networks with BC and MAC, demonstrating optimality in trees.

Findings

Algorithm achieves near-optimal throughput in tree networks.

Exploiting BC and MAC improves multihop network performance.

Multiuser local pooling factor characterizes algorithm performance.

Abstract

Information theoretic Broadcast Channels (BC) and Multiple Access Channels (MAC) enable a single node to transmit data simultaneously to multiple nodes, and multiple nodes to transmit data simultaneously to a single node respectively. In this paper, we address the problem of link scheduling in multihop wireless networks containing nodes with BC and MAC capabilities. We first propose an interference model that extends protocol interference models, originally designed for point to point channels, to include the possibility of BC and MAC. Due to the high complexity of optimal link schedulers, we introduce the Multiuser Greedy Maximum Weight algorithm for link scheduling in multihop wireless networks containing BCs and MACs. Given a network graph, we develop new local pooling conditions and show that the performance of our algorithm can be fully characterized using the associated parameter, the multiuser local pooling factor. We provide examples of some network graphs, on which we apply local pooling conditions and derive the multiuser local pooling factor. We prove optimality of our algorithm in tree networks and show that the exploitation of BCs and MACs improve the throughput performance considerably in multihop wireless networks.