Clique-Based Utility Maximization in Wireless Mesh Networks

Abstract

This study considers utility-based resource allocation in backbone wireless mesh networks (WMNs). Unlike single-hop cellular networks, a WMN has multi-hop transmissions with multiple contending links, and thus requires more careful design for resource allocation. To this end, we provide a clique-based method with efficient spatial reuse, which is then incorporated into proportionally fair scheduling (PFS) for fair resource management in WMNs. We call it a clique-based proportionally fair scheduling (CBPFS) algorithm. The linear and/or logarithmic rate models used to analyze PFS in single-hop cellular networks cannot be used to analyze CBPFS in backbone WMNs. Using stochastic approximation and recent results on rate modeling for Rayleigh fading channels, we conduct mathematical analysis and obtain a closed-form model to quantify CBPFS performance, without the need of the highly time-consuming ordinary differential equation (ODE) analysis. We use the derived analytical framework to estimate the link throughput of CBPFS and compare it with simulations. It is the first time a closed-form analytical model is developed to quantify the throughput of links in a multi-hop network where links are proportionally fair scheduled.