Robust interference management via outage-constrained downlink beamforming in multicell networks

We introduce a downlink robust optimization problem to minimize an objective function which is a combination of a cost characterizing the total transmit power at each multiple antenna base station (BS) and a penalty term due to the induced aggregate intercell interference power across the multicell network. This optimization is constrained to assure that a set of target signal-to-interference-plus-noise ratio (SINR) levels are maintained at the intracell user terminals with adjustable outage probabilities. The involved uncertainties are due to the errors incurred in estimating both direct downlink channels within each cell and the interfering channels between a BS and the remote user terminals in other cells. By introducing a slack variable, we reformulate the problem so that the probability of imperfection in the penalty term is confined within an adjustable outage. To maintain the tractability of the robust solutions, we derive an equivalent semidefinite programming (SDP) formulation that is convex under the standard rank relaxation. Our simulation results show that at fixed outage probabilities, the proposed scheme is considerably more sensitive to imperfection in direct downlink channels than the imperfection in the interfering channels at a given BS. This observation is due to the presence of the interfering channels within the penalty term that is minimized as a part of the proposed objective function.