ID	原文	译文
3043	从基站分簇、用户分组、资源块分配与功率分配方面共同优化系统能效，最大限度地削弱基站簇间干扰与簇内干扰。	Jointly optimizesystem energy efficiency in terms of base station clustering, user grouping, resource block allocation and power alloca-tion, which minimized the inter-cluster interference and intra-cluster interference of the base station efficiently.
3044	仿真结果表明，所提算法在能效和计算效率相较对比算法均有明显优化。	The sim-ulation results show that the proposed algorithm is better on energy efficiency and computational efficiency comparedwith existing algorithms.
3045	为了解决传统基于最优化方法所设计的无线网络资源管理策略通常复杂度较高且实时性差，不利于在线决策制定的问题，针对基于 SWIPT 的传感云系统，建立汇聚（sink）节点能效最大化问题及其数学模型，	To solve the problems of high complexity and poor real-time performance caused by traditional wireless re-source management based on optimization methods, the energy efficiency maximization problem of sink node and its mathematical model were established for SWIPT-enabled sensor-cloud system,
3046	然后引入深度学习方法，通过对最优化算法的学习实现更低复杂度与更高实时性的算法设计。	then the deep learning method was intro-duced to realize the solving and online decision-making with lower complexity and higher real-time performance.
3047	为了实现深度学习算法在网络资源分配中的应用，首先将 sink 节点最优能效模型转化为高维可求解形式，设计具有迭代形式的SWIPT-WMMSE 算法实现最优波束成形矢量的求解，同时证明了算法的收敛性。	The mathematical model was transformed into a high-dimensional solvable form, and then a SWIFT-WMMSE algorithm withiterated forms was designed to solve optimal beamforming vector. The convergence of SWIPT-WMMSE algorithm wasproved.
3048	然后基于 DNN 逼近误差的传递过程推导了 DNN 设计准则，并通过对 DNN 的训练实现其对 SWIPT-WMMSE 算法的逼近。	Then, based on error propagation of DNN approximation, the scale design criteria for the DNN was deduced, andthe approximation was realized through DNN training.
3049	最后通过仿真实验分别验证了 SWIPT-WMMSE 算法与 DNN 算法的有效性，及 DNN 算法的逼近效果和在提升系统性能方面的优势。	Finally, the simulation results verify the effectiveness of SWIPT-WMMSE and DNN algorithm, as well as the approximation effect of DNN and its system performance gains.
3050	针对认知容量收集网络中网络切片的频谱共享问题，提出了一种认知容量收集网络中网络切片频谱共享策略。	To realizing the network slice with spectrum sharing in cognitive capacity harvesting network, a spectrum sharing strategy was proposed based on 4D conflict graph and opportunistic capacity.
3051	通过建立 4D 冲突图模型，提出了无冲突节点集的求解方法，	Firstly, a 4D conflict graph modelfor mesh network was built and a method to achieve the node sets with conflict free was proposed.
3052	并建立了频谱共享的机会容量模型，推导了非授权信道的机会容量，联合 4D 冲突图模型和机会容量模型提出了一种认知容量收集网络中网络切片频谱共享策略。	Then, the opportunis-tic capacity model of spectrum sharing was established and the opportunistic capacity of the unlicensed channel was de-rived. Finally, the spectrum sharing strategy based on 4D conflict graph and opportunistic capacity was proposed for cog-nitive capacity harvesting network.