Joint Optimization of Beamforming and Power Allocation for Multicell Downlink Systems

Abstract

This paper investigates a joint optimization design method for beamforming and power allocation to improve the energy efficiency (EE) of a multicell MIMO downlink system considering the effect of channel estimation errors. The channel state information (CSI) imperfections can be well modeled with the power for training sequence. Assuming that both base stations and users are equipped with multiple antennas, this research focuses on the joint optimization of the beamforming vectors and the power allocation ratio between training sequence power and transmit data in order to maximize the EE under power constraints. A robust alternating optimization based on iterative algorithm to solve this problem which is in a nonconvex fractional form is proposed. First, the fractional problem is transformed into a linear form using the Dinkelbach method. The sumrate maximization problem is, then, replaced by the sum-MSE minimization problem by applying weighted mean square error minimization (WMMSE) method. Finally, the expectation taken over the distribution of the channels can be approximated using the sample average approximation (SAA) and the problem can be solved by computing a second order cone programming (SOCP). The robustness and effectiveness of the proposed method are validated by the simulation results.