Link Scheduling for Throughput Maximization in Multihop Wireless Networks Under Physical Interference

TL;DR

This paper develops efficient link scheduling algorithms for multihop wireless networks under a realistic physical interference model, achieving near-optimal throughput with approximation ratios independent of network size.

Contribution

It introduces novel scheduling algorithms that connect physical and graph-based interference models, providing approximation guarantees for different power adjustment scenarios.

Findings

Achieves O(g(E)) approximation for dynamic power adjustment.

Provides O(g(E)) and O(g(E)logρ) approximations for fixed power scenarios.

Approximation factors are independent of network size.

Abstract

We consider the problem of link scheduling for throughput maximization in multihop wireless networks. Majority of previous methods are restricted to graph-based interference models. In this paper we study the link scheduling problem using a more realistic physical interference model. Through some key observations about this model, we develop efficient link scheduling algorithms by exploiting the intrinsic connections between the physical interference model and the graph-based interference model. For one variant of the problem where each node can dynamically adjust its transmission power, we design a scheduling method with O(g(E)) approximation to the optimal throughput capacity where g(E) denotes length diversity. For the other variant where each node has a fixed but possible different transmission powers for different nodes, we design a method with O(g(E))-approximation ratio when the transmission powers of all nodes are within a constant factor of each other, and in general with an approximation ratio of O(g(E)log\rho) where log\rho is power diversity. We further prove that our algorithm for fixed transmission power case retains O(g(E)) approximation for any length-monotone, sub-linear fixed power setting. Furthermore, all these approximation factors are independent of network size.