Decentralized multiuser beamforming for cellular communication systems

This paper explores multiuser downlink beamforming in a distributed manner, where each user locally and sequentially optimizes its own beamforming phase and amplitude according to some performance measures. For phase optimization, the complex non-convex optimization problem is reformulated into a convex one. Given the optimal phase, the amplitude is pursued through the method of Lagrange multiplier. Our numerical simulations demonstrate that the proposed algorithm matches the conventional centralized two-user zero-forcing beamforming performance in terms of bit error rate. Meanwhile, our method utilizes 50% less feedback as compared to the conventional one. As a result, a low complexity distributed multiuser beamforming is delivered.