Joint admission control and beamforming in max–min fairness networks

Abstract

The max–min fairness (MMF) strategy has been widely employed to manage wireless networks since it guarantees fairness among users. However, with a large number of users awaiting service, the network tends to be congested and the quality-of-service (QoS) will degrade substantially. This motivates us to study the MMF problem jointly with the consideration of admission control. Specifically, the authors consider a downlink network consisting of a multi-antenna base station (BS) and multiple single-antenna users. By jointly optimising the admissible users and the BS transmit beamformers, they aim to maximise the minimum signal-to-interference-plus-noise-ratio of the admissible users, such that high QoS and fairness can be guaranteed simultaneously for them. This problem is essentially NP-hard, and hence they pursue an efficient approximate solution to it. To this end, they first reformulate this problem from the perspective of sparse optimisation, and then develop a low-complexity algorithm to iteratively solve the approximate problem. Moreover, to facilitate the algorithm's implementation, they further recast the subproblem in each iteration, such that it fits into the framework of the alternating direction methods of multipliers. Finally, an efficient distributed algorithm is designed, with each step being simply computed in a closed form.