通信侧与供电侧双侧随机的Massive MIMO超密集异构网络资源分配研究

In 5G system, network densification exacerbates the number of network nodes and the substantial increment of antennas aggravates the processing overhead of system signal. It has become an urgent issue that balancing network performance and controllable system loss which includes power consumption and energy cost. So the diversity of power supply provides opportunities of energy conservation and cost reduction. Considering the randomness of power requirement for base station caused by the demand difference of user QoS, user allocation and interference suppression in communication side, considering the randomness, intermittence and volatility of renewable energy in supply side, an integrated framework for Massive MIMO ultra-dense heterogeneous network with double random character of communication side and supply side is constructed in this research. And a tradeoff between economics of system power supply and environmental costs, a resource allocation optimization mechanism under the 5G network and smart grid background, is also investigated in this research. More deeply, to satisfy the need of high performance, reasonable complexity and controllable energy cost, the resource allocation optimization schemes will consider the factors of the system performance, power supply, energy cost and the user fairness etc., will make a full use of various resources including space domain, time domain, frequency domain and energy domain, will use the theory of big data prediction, game theory, multi-objective optimization etc.. The resource allocation scheme for Massive MIMO ultra-dense heterogeneous network will provide a necessary theory basis to achieve the goal of improving cost efficiency and reducing carbon emission in 5G.

5G系统中网络的密集化使得网络节点数量更加庞大，大规模天线的使用增加了系统信号处理开销，在实现网络性能的同时控制系统总能耗和供电成本成为重要的课题。而基站供电方式的多样性，为能量节约和成本降低提供新的机遇。为此，课题拟将综合考虑通信侧用户QoS差异、用户分配以及干扰抑制引起的基站功率需求的随机性，以及可再生能源具有随机性、间歇性、波动性的特点，构建通信侧与供电侧双侧随机的Massive MIMO超密集异构网络一体化框架，权衡系统供电的经济和环境代价，探讨5G网络和智能电网背景下的资源分配优化机制。综合考虑系统性能、供电方式、能源代价、用户公平性等因素，充分利用空域、时域、频域和能量域资源，运用大数据分析、博弈论、多目标优化等方法，探索形成系统性能优良、计算复杂度合理、能量代价可控的Massive MIMO超密集异构网络资源分配方案，为实现5G系统提高成本效率、降低碳排放的愿景提供理论支撑。

5G系统中网络的密集化使得网络节点数量更加庞大，大规模天线的使用增加了系统信号处理开销，在实现网络性能的同时控制系统总能耗和供电成本成为重要的课题。而基站供电方式的多样性，为能量节约和成本降低提供新的机遇。 . 项目组综合考虑通信侧用户QoS差异、用户分配以及干扰抑制引起的基站功率需求的随机性，以及可再生能源具有的随机性、间歇性、波动性特点，结合多点协作技术和智能电网技术，开展了超密集异构网络无线通信侧和供电侧的双侧随机性研究。提出了Massive MIMO信道信息获取新方法以及用户移动时的导频分配算法；开展了多时间尺度下综合能源系统低碳优化调度研究；提出了超密集部署下绿色能源感知的用户关联方法。构建了基于多源供电的多基站协作系统模型，建立了基站系统功率和能源损耗模型，模拟了太阳能和风能的产能速率，采取可再生能源与传统能源混合供能方式为基站系统提供能量。建立了能量代价效率函数，理论证明了代价效率函数存在全局最优的解，且该解是系统最优的代价效率。以降低能量代价为目标，综合考虑系统性能、供电方式等因素，充分利用空域、时域、频域和能量域资源，形成了基于能量代价的多维联合的资源分配方案，提出了基于匹配博弈和凸优化的混合供能算法，由线性规划和多目标优化组成的混合优化算法以及超密集异构网中的Q学习资源调度算法等资源分配方案，为实现5G系统提高成本效率、降低碳排放的愿景提供了理论支撑。