Convex optimization for joint zero-forcing and antenna selection in multiuser MISO systems

The problem of joint zero-forcing (ZF) beamforming (BF) together with optimal power allocation (PA) and antenna selection (AS) for throughput maximization is considered in this paper for multi-user multiple input single output (MU-MISO) systems. We introduce a new formulation for the joint ZF and PA problem by adapting the algebraic subspace approach which finds a proper set for the optimization variable that inherently satisfies the ZF constraints. Also, the squared group Lasso penalty on the BF matrix is used to linearize (relax) the non-convex, NP-hard problem of joint BF and AS. Extensive simulations show that for the throughput problem, the proposed algorithm performs very closely to the optimal (exhaustive search) joint approach.