Reduced-complexity power allocation in zeroforcing MIMO-OFDM downlink system with multiuser diversity

This paper considers the power allocation for a multiuser MIMO-OFDM downlink system using zero-forcing multiplexing, with the objective to maximize the system capacity. An optimal power allocation algorithm has been developed and is presented. While the optimal power allocation algorithm achieves promising capacity, its complexity increases exponentially with the number of users. We develop two power allocation algorithms whose complexities increase only linearly with the number of users, with only negligible loss of capacity compared to the optimal algorithm. Numerical results and analysis are provided to evaluate the system capacities of the optimal and the reduced-complexity algorithms. At SNR of 10 dB, the optimal and the suboptimal schemes attain 40% capacity gain over a round-robin scheme with water filling and 60% capacity gain over a round-robin scheme with equal power allocation