Joint Admission Control and Beamformer Design for Mobile Users: Stay Here or Move to a Better Position?

Abstract

In this paper, we study the joint admission control and beamforming problem within a network where one multi-antenna base station tries to serve multiple single-antenna users. Unlike most existing studies which merely identify the users that should be denied, our work further suggest better positions in the neighbouring area for them where the previously-rejected users are allowed to access the network. To address this, we assume the knowledge of channel vectors within the network coverage area, and then jointly optimize the transmit beamformer and the channel vector associated with each user to minimize the network power cost, with a penalty measuring the mismatch between the optimized channel and the reference channel with current user position. Specifically, a non-zero mismatch means that the corresponding user is inadmissible at present, but may access network if moving to the position with the channel closest to the optimized result. Basically, this is a challenging non-convex problem, and we design a penalty dual decomposition (PDD)-based algorithm to iteratively achieve a stationary solution. The algorithm is highly efficient since a simple analytical solution is derived in each step.