ID 原文 译文
52537 为有效提升多输入多输出系统的频谱效率，降低算法复杂度，文章提出了一种基于迭代思想的矩阵分解算法。 In order to effectively improve the spectrum efficiency of the Multiple-Input Multiple-Output (MIMO) system and reduce the complexity of the algorithm, this paper proposes a matrix decomposition algorithm based on iterative method.
52538 首先初始化模拟预编码矩阵并获取最优预编码矩阵; First, the analog precoding matrix is initialized and the optimal precoding matrix is obtained.
52539 然后通过模拟预编码矩阵与最优预编码矩阵来获取数字预编码矩阵，通过提取数字预编码矩阵中的相位信息更新模拟预编码矩阵，利用已更新的模拟矩阵来更新数字矩阵; Then the digital precoding matrix is obtained by simulating the precoding matrix and the optimal precoding matrix, and the analog precoding matrix is updated by extracting the phase information in the digital precoding matrix.
52540 最后通过不断交替迭代直到收敛。 The updated simulation matrix updates the digital matrix and finally iterates through alternating iterations until convergence.
52541 仿真结果表明，所提算法具有较低的复杂度且更接近纯数字预编码性能。 The simulation results show that it has lower complexity and better performance, which is closer to the pure digital precoding performance.
52542 为实现光保真技术-无线保真技术 (LiFi-WiFi) 混合网络的吞吐量最优化，文中将系统负载均衡问题建模为联合效用最大化问题， Aiming at throughput optimization of hybrid LiFi-WiFi network, the system load balancing scheme can be formulated as a joint utility maximization problem,
52543 并使用对数效用函数实现用户的比例公平。 and a logarithmic utility function is used to achieve proportional fairness for users.
52544 在该方案中考虑LiFi和WiFi网络之间切换导致的信令开销，结合LiFi系统直射和反射路径以及信道噪声，分析了光同信道干扰和非同信道干扰情况下信噪比对数据速率和总吞吐量的影响。 In this scheme, based on the signalling overhead caused by handover between light-fidelity (LiFi) and wireless fidelity (WiFi) , as well as the channel noise from the direct and reflective paths, the data rate and total throughput effected by SNR are analyzed and evaluated with co-channel interference and non-channel interference.
52545 仿真结果表明，该算法的估计值满足实际的通信需求。 The simulation results show that the estimation of the proposed scheme can meet the actual communication demand.
52546 针对现有的基于互信息最大化的异构图神经网络(HGNN)方法因图读出操作的单射限制、粗粒度的特征保留而无法适用于现实网络的问题，提出一种基于局部图互信息最大化的、无监督的异构图神经网络方法。 Aiming at the shortcomings of the injective ability of readout function and coarse-grained feature preservation in traditional mutual information maximization based heterogeneous graph neural networks( HGNN), which make them inadequate to use in the real-work networks, a new local graphical mutual information maximization based unsupervised heterogeneous graph neural network is presented.