Joint Transmit Beamforming and Antenna Selection for Energy Efficiency Maximization in MISO Downlink

We study the joint beamforming and antenna selection problem for energy efficiency maximization in multi-user multiple-input single-output (MISO) downlink channel. By viewing antenna selection as finding a sparse solution, we first introduce a sparsity-inducing regularization term to the design problem. Since the resulting problem is nonconvex, it is difficult to find an optimal solution, and we apply a local optimization method based on the concept of sequential convex approximation (SCA) to solve this problem. By proper reformulations we arrive at a fast converging iterative algorithm, where a convex program is solved at each iteration. In the first design, we simply ignore antennas of which the associated beamformers are nearly zero and select the remaining ones. In the second design, we further perform the search over the selected antennas of the first design to improve the energy efficiency. Numerical results demonstrate remarkable performance gains of the proposed approaches in terms of energy efficiency over the solution without antenna selection.