Transmit Design for Massive MIMO Multicasting With Statistical CSIT

We investigate physical layer massive multiple-input-multiple-output (MIMO) multicasting transmit design with statistical channel state information at the base station. We first establish the relationship between the transmit design problems under the quality of service and max-min fair criteria. Then we focus on the transmit designs under the latter criterion. We show that the eigenvectors of optimal input covariance are given by the columns of the discrete Fourier transform matrix for the uniform linear array, which reveals the optimality of beam domain transmission in massive MIMO multicasting. We further propose a dual algorithm together with stochastic programming to specify the eigenvalues of input covariance. In addition, a simplified input covariance optimization by applying the deterministic equivalent technique is presented to reduce the complexity involved in stochastic programming. Simulation results demonstrate the performance of the proposed algorithms.