We investigate the max-min weighted downlink signal-to-interference ratio (SINR) problem under uplink SINR
constraints and practical per-antenna constraints in full-duplex systems. The successive convex approximation (SCA)
method is adopted to iteratively deal with this non-convex problem. Within each SCA iteration, to lower the complexity, a parallel beamforming algorithm based on alternating direction method of multipliers (ADMM) is proposed. Specifically, local variables are introduced to decompose the problem to multiple independent subproblems with closed-form solutions. Numerical results show that our proposed algorithm can achieve the similar performance with existing algorithms, but runs much faster especially in large-scale systems.