Optimal joint base station association and beamforming design for downlink transmission

This paper presents a solution framework to jointly optimize the base station association strategy and linear beamforming design for downlink transmission in a multicell system. Assuming each mobile station can only be assigned to one base station, our design objective is to minimize the sum transmit power across the base stations with a set of target signal-to-interference-plus-noise ratios at the mobile stations. Since the original optimization problem involves binary variables for base station associations, finding its optimal solution is a challenging task. To circumvent this difficulty, the original problem is relaxed into a new optimization problem by expanding its constraint set. Interestingly, it is shown that the relaxed problem can be solved optimally and its solution is also optimal to the original problem. We then propose two solution approaches to tackle the relaxed problem: one via its Lagrangian dual problem and the other via its dual uplink problem. Simulation results show that the resulting jointly optimal base station association and beamforming design can significantly outperform fixed base station association schemes.