ID 原文 译文
17725 但是由于MPD算法的计算复杂度随调制阶数和用户天线数的增加而增加，而概率近似消息传递检测(PA-MPD)算法可以减少MPD算法的计算复杂度。 However, the computationalcomplexity of the MPD algorithm increases with the increase of the modulation order and the number of userantennas, and the Probability Approximation Message Passing Detection (PA-MPD) algorithm can reduce thecomputational complexity of the MPD algorithm.
17726 为了进一步降低PA-MPD算法的复杂度，该文在PA-MPD算法的基础上引入了提前终止迭代策略，提出了一种改进的概率近似消息传递检测算法(IPA-MPD)。 In order to further reduce the complexity of PA-MPDalgorithm, this paper introduces an early termination iteration strategy based on PA-MPD algorithm, andproposes an Improved PA-MPD (IPA-MPD) algorithm.
17727 首先确定不同用户的符号概率在迭代过程中的收敛速率，然后根据收敛率来判断用户的符号概率是否达到最佳收敛，最后对符号概率到达最佳收敛的用户终止算法迭代。 Firstly, the convergence rate of the symbol probability of different users in the iterative process is determined, and then the convergence probability is used to determine whether the user’s symbol probability reaches the best convergence. Finally, the user terminationalgorithm that the symbol probability reaches the best convergence is iterated.
17728 仿真结果表明，在不同单天线用户配置下IPA-MPD算法的计算复杂度可降低为PA-MPD算法的52%～77%，且不损失算法的检测性能。 The simulation results showthat the computational complexity of the IPA-MPD algorithm can be reduced to 52%～77% of the PA-MPDalgorithm under different single-antenna user configurations without loss of the detection performance of thealgorithm.
17729 车载云计算环境中的计算卸载存在回程网络延迟高、远程云端负载大等问题， In the vehicular cloud computing environments, computation offloading faces the problems such ashigh network delay and large load of the remote cloud.
17730 车载边缘计算利用边缘服务器靠近车载终端，就近提供云计算服务的特点，在一定程度上解决了上述问题。 The vehicular edge computing takes advantage of theedge servers to be close to the vehicular terminals, and provides the cloud computing service to solve theproblem mentioned above.
17731 但由于汽车运动造成的通信环境动态变化进而导致任务完成时间增加，为此该文提出一种基于移动路径可预测的计算卸载切换策略MPOHS，即在车辆移动路径可预测情况下，引入基于最小完成时间的计算切换策略，以降低车辆移动性对计算卸载的影响。 However, due to the dynamic change of communication environment caused by vehicle movement, the task completion time will increase. For this reason, this paper proposes a MobilityPrediction-based computation Offloading Handoff Strategy (MPOHS), which tries to minimize the average completion time of offloaded tasks by migrating tasks among edge servers according to the prediction of vehicle movement.
17732 实验结果表明，相对于现有研究，该文所提算法能够在减少平均任务完成时间的同时，减少切换次数和切换时间开销，有效降低汽车运动对计算卸载的影响。 The experimental results show that, compared with the existing research, the proposed strategy can reduce the average task completion time, cut down the handoff times and handoff time overhead, and effectively reduce the impact of vehicle movement on the performance of computation offloading.
17733 稀疏多元逻辑回归(SMLR)作为一种广义的线性模型被广泛地应用于各种多分类任务场景中。 As a generalized linear model, Sparse Multinomial Logistic Regression (SMLR) is widely used invarious multi-class task scenarios.
17734 SMLR通过将拉普拉斯先验引入多元逻辑回归(MLR)中使其解具有稀疏性，这使得该分类器可以在进行分类的过程中嵌入特征选择。 SMLR introduces Laplace priori into Multinomial Logistic Regression (MLR)to make its solution sparse, which allows the classifier to embed feature selection in the process of classification.