Utility-proportional fairness in wireless networks

Abstract

Current communication networks support a variety of applications with different quality of service (QoS) requirements which compete for its resources. This continuously increasing competition highlights the necessity for more efficient and fair resource allocation. Current Network Utility Maximization (NUM) framework fails to achieve this target and alternative approaches cannot operate in networks that consist of wireless links. This paper presents a NUM framework for wireless networks that shares resources according to the utility proportional fairness policy. This policy is shown to prevent rate oscillations in the resource allocation process, allocate resources in a more fair manner among different types of applications and lead to the calculation of closed form solutions for the optimal rate allocation function. Based on this policy, a distributed rate and power allocation algorithm is proposed that gives priority to applications with greater need of resources. Finally, numerical results on the performance of the proposed algorithm are presented and compared against other approaches in the literature.