Energy efficiency optimization in training-based multicell coordinated beamforming systems

In this paper, power allocation between training and data symbols is investigated to maximize energy efficiency (EE) for multicell systems using coordinated beamforming (CBF). We formulate the EE optimization problem under a minimum rate requirement as a multi-variable non-quasiconcave fractional program. Based on an approximation of the ergodic rate, we transform the problem into a jointly quasiconcave fractional program and propose a numerical method to solve it in two steps. Firstly, an alternating optimization algorithm is proposed to solve it without the minimum rate requirement. Secondly, we need to judge whether the solutions satisfy this requirement. If not, a problem of minimizing the total power consumption is reformulated and we propose a linear programming algorithm to solve it. Simulation results show that our method can converge within a few iterations and achieve a better EE performance compared to the spectrum-efficient scheme.