全局最优的SLNR预编码功率分配算法

针对多用户多输入多输出系统,基于总可达速率最大化原则,研究了基于信漏噪比(signal-to-leakage-and-noise ratio, SLNR)预编码功率分配算法。为解决传统算法仅能通过迭代求解,性能不好的问题,提出了可直接求得全局最优解的解决方案。导出了仅含一种变量的目标函数,将功率分配问题转化为有约束条件的非凸优化问题。为解决这个优化问题,提出一种基于鸟群算法(bird swarm algorithm, BSA)实现全局最优的解决方案;为弥补该算法复杂度高的缺点,提出一种基于几何规划(geometric programming, GP)的近似全局最优算法。仿真表明,基于BSA的算法能够实现最大的系统总速率,且随着发射信噪比、基站天线数及用户数变化,性能提升效果能够得到保持;基于GP的算法,在复杂度与性能方面得到折衷,且相比基于BSA的算法,误码率性能更优。

Global optimal power allocation algorithm for SLNR precoding

Based on the principle of maximizing sum achievable rate, power allocation algorithms for SLNR precoding are investigated for the multi-user multiple-input multiple-output (MU-MIMO) system. To solve the problem that the solutions are not of good quality with traditional algorithms because they can only be obtained in an iteration way, a global optimal scheme is proposed to get the solution directly. Firstly, an objective function containing only one set of variables is obtained by derivation, which converts the power allocation problem to a non-convex optimization problem with constraint conditions. To solve the optimization problem, a scheme based on bird swarm algorithm (BSA) achieving global optimization is proposed. To compensate the disadvantage of the former algorithm in which the complexity is high, a scheme based on geometric programming (GP) achieving approximated global optimization is proposed. According to the simulation, the algorithm based on BSA achieves the highest sum rate, and the effect of performance improvement can be maintained with change of transmit signal-to-noise ratio, the number of base station antennas and users. The algorithm based on GP makes compromises between complexity and performance. Furthermore, compared with the algorithm based on BSA, it can get better bit error rate (BER) performance.