Rate Constrained Decentralized Beamforming for MIMO Interfering Broadcast Channel

We consider rate constrained decentralized downlink transceiver optimization for cellular systems. Our optimization objective is the weighted sum rate, which provides generalized objective for wide class of priority and throughput management problems. The quality-of-service imposing rate constraints make the problem considerably more complex, and have been largely neglected in the past in the context of utility maximization. The decentralized design is achieved with a novel combination of successive convex approximation and alternating direction method of multipliers (ADMM). While providing more sophisticated interference management, the considered algorithm does not increase the computational complexity, when compared to other recently published rate optimizing methods. The proposed design is shown, by numerical evaluation, to provide significant improvement in the convergence properties. Furthermore, we show that the ADMM approach provides much needed stability in the time correlated channels.