中文摘要：

小区之间优化变量耦合使多小区联合优化问题为非线性非凸问题,通过约束泄露干扰提出了基于半正定松弛的中心式协同波束成形算法,在此基础上,采用部分对偶分解方法将中心式联合优化问题分解为一组单小区优化问题,提出了一种分布式迭代求解算法。结果表明,所提算法在保证系统安全速率需求和认知干扰约束的条件下实现了系统功耗的降低,此外,分布式算法不仅能降低实现复杂度,而且能达到中心式算法的性能。

英文摘要：

Owing to the coupling among the vectors to be optimized, the problem becomes non-convex and non-linear. Accordingly, a centralized coordinated beamforming design method was proposed by constraining the leakage interfer- ence and using the semi-definite relaxation, and on the base of it, a decentralized algorithm was also proposed by de- composing the centralized design problem into a group of sub-problems on each cell＇s base by using dual decomposition. Simulation results show that the proposed schemes make effective reduction in power consumption while guarantee the legal users＇ secrecy rate requirement and the cognitive interference constraint. Moreover, the proposed decentralized method not only reduces the computational complexity, but also has the same performance as the centralized algorithm.