Computational efficient transceiver optimization for multiuser MIMO systems: Power minimization with user-MMSE requirements

This paper studies transceiver design for multi-user MIMO systems and focuses on the problem of minimizing the total transmit power while maintaining individual minimum mean square error (MMSE) requirement per user. We propose a new iterative algorithm and prove that the total transmit power decreases monotonically and converges to an optimum. The algorithm has low complexity and fast convergence behavior during the first few iterations. The main contribution is a new power allocation strategy that achieves per-user MMSE targets. The optimal powers can be obtained by solving a linear system. This results in much lower complexity as compared with the strategy proposed in [1], where a fixed point iteration is required.