ID 原文 译文
25745 针对 LEO 卫星网络在多跳转发数据包时流量分布不均问题，提出了一种基于不完全信息的最优收益路由联盟博弈算法。 To tackle the difficulty of imbalanced traffic load when forwarding data packets through multiple hops in LEO satellite network, we propose a collaborative game theory routing strategy with incomplete information．
25746 各节点协同联盟邻居节点，共同确定数据报文当前最优转发路径，从而分配和平衡节点间流量负载。 Each node determines the current optimal routing path of the data packets through cooperating with neighbors for locally optimized coalition gain. This strategy distributes and balances the traffic workload among neighbor nodes effectively．
25747 仿真结果表明，与最短路径卫星路由 DSP 或智能路由 TLＲ 相比，本文算法的平均数据传输延迟降低了 18. 5% ，节点流量负载均衡度提高了 65. 6% 。 Simulations show that compared with satellite routing based on the DSP shortest path or the TLＲ intelligent routing, the proposed algorithm reduces the average transmission delay by 18. 5% , and improves the load balancing of nodes by 65. 6% ．
25748 相较于传统正交多址接入，非正交多址接入技术由于在系统吞吐量、频谱效率和能量效率等方面的优势，使其成为 5G 多址技术研究热点。 As non-orthogonal multiple access technology can achieve higher system throughput，spectrum efficiency and energy efficiency than traditional orthogonal access technology，it has become a research hotspot of 5G multiple access technology．
25749 针对 NOMA 下行链路的系统能量效率优化问题，提出一种基于改进粒子群算法的功率分配策略。 In this paper, a power allocation strategy based on Improved Particle Swarm Optimization ( IPSO) is proposed to optimize the energy efficiency of NOMA downlink system．
25750 建立了基于能量效率最大化的优化模型，在标准粒子群算法的基础上提出三点改进，并将改进后的粒子群算法用于求解最大化系统能效的目标函数。 The Standard Particle Swarm Optimization ( SPSO) is improved in three aspects, and the IPSO algorithm is used to solve the objective function to maximize the energy efficiency of the system．
25751 研究结果表明，在最佳功率分配点处，改进后的粒子群算法使系统能量效率显著提高。 The simulation results show that at the optimal power allocation point, the IPSO algorithm can significantly improve the energy efficiency of the system．
25752 本文提出了一种基于改进 Wasserstein 生成式对抗网络( De-aliasing Wasserstein Generative AdversarialNetwork with Gradient Penalty，DAWGAN-GP) 的磁共振图像重构算法，该方法利用 Wasserstein 生成式对抗网络代替传统的生成式对抗网络，并结合梯度惩罚的方法提高训练速度，解决 WGAN 收敛缓慢问题。 In this paper, we propose an improved Wasserstein generative adversarial network ( WGAN) , de-aliasing Wasserstein generative adversarial network with Gradient Penalty ( DAWGAN-GP) , for magnetic resonance imaging ( MＲI) reconstruction. This method uses WGAN to replace the traditional GAN, and combined the gradient penalty to improve thetraining speed and to solve the slow convergence problem of WGAN.
25753 此外，为了有更好的重构效果，我们将感知损失，像素损失和频域损失引入至损失函数中进行网络训练。 In addition, for better preservation of the fine structuresin the reconstructed images, we incorporate perceptual loss, pixel loss and frequency loss into the loss function for training the network.
25754 实验结果表明，对比现有的基于深度学习的磁共振图像重构算法，基于 DAWGAN-GP 的磁共振图像重构方法具有更好的重构效果，可获得更高的峰值信噪比( Peak Signal to Noise Ｒatio，PSNＲ) 和更好的结构相似性( Structural Similarity Index Measure，SSIM) 。 Compared with other state-of-the-art deep learning methods for MＲ images reconstruction, DAWGAN-GPmethod outperforms all other methods and can provide superior reconstruction with improved peak signal to noise ratio( PSNＲ) and better structural similarity index measure ( SSIM) ．