基于块循环矩阵的对称张量的最佳秩-1逼近

对称张量的最佳秩-1问题是张量研究中非常重要的部分.首先,基于三阶张量的块循环矩阵,提出了求解对称张量最佳秩-1逼近问题的一个新方法.其次,针对求解对称张量的最佳秩-1逼近方法,给出了对称张量的最佳秩-1逼近不变性的一个充要条件,以及逼近误差上界的估计.最后,数值算例表明了上述方法的可行性和误差上界的正确性.     

负载均衡问题

自Ron Graham20世纪60年代发表第一篇负载均衡算法的论文以来,平行机排序作为组合优化近似算法理论的首个问题引起了学界的广泛兴趣,其本身研究的不断深化也一路见证了该领域的发展历程.介绍负载均衡问题的来龙去脉,特别是作者所在团队在相关问题的研究进展,从算法和复杂性不同的视角分析经典问题的计算本质,并对未来的研究提出一些建议.     

四阶弱对称非负张量Z-谱半径的上下界及应用

针对四阶张量Z-谱半径的估计问题,利用张量Z-特征值的定义,并结合不等式放缩技巧,给出了四阶弱对称非负张量Z-谱半径的新上下界,改进了现有一些结果.作为应用,由Z-谱半径的上界给出了张量最佳秩一逼近和贪婪秩一更新算法收敛速度的下界,由Z-谱半径的上下界给出了具有非负振幅对称纯态纠缠的几何度量的上下界.     

A new approximation for the generalized fractional derivative and its application to generalized fractional diffusion equation

In this paper, we derive a fourth order approximation for the generalized fractional derivative that is characterized by a scale function z(t) and a weight function w(t) . Combining the new approximation with compact finite difference method, we develop a numerical scheme for a generalized fractional diffusion problem. The stability and convergence of the numerical scheme are proved by the energy method, and it is shown that the temporal and spatial convergence orders are both 4. Several numerical experiments are provided to illustrate the efficiency of our scheme.     

On Poincaré-Friedrichs Type Inequalities for the Broken Sobolev Space ${\rm W}^{2,1}$

We are concerned with the derivation of Poincaré-Friedrichs type inequalities in the broken Sobolev space $W^{2,1}$($Ω$; $\mathcal{T}_h$) with respect to a geometrically conforming, simplicial triagulation $\mathcal{T}_h$ of a bounded Lipschitz domain $Ω$ in $\mathbb{R}^d$ , $d$ $∈$ $\mathbb{N}$. Such inequalities are of interest in the numerical analysis of nonconforming finite element discretizations such as ${\rm C}^0$ Discontinuous Galerkin (${\rm C}^0$${\rm DG}$) approximations of minimization problems in the Sobolev space $W^{2,1}$($Ω$), or more generally, in the Banach space $BV^2$($Ω$) of functions of bounded second order total variation. As an application, we consider a ${\rm C}^0$${\rm DG}$ approximation of a minimization problem in$BV^2$($Ω$) which is useful for texture analysis and management in image restoration.     

带有输入饱和的TCP/AQM网络系统事件触发控制

研究了具有用户数据报协议(user datagram protocol,UDP)数据流干扰和不确定项的TCP/AQM网络拥塞控制问题.首先考虑由丢包率范围而导致的输入饱和的问题,从而对丢包率进行限制.其次,利用模糊逼近的方法处理系统中的干扰和不确定项.为节约网络资源,引入事件触发机制,结合输入饱和、模糊逼近以及Backstepping技术,提出了一种主动队列管理算法,不仅可以使队列长度跟踪到期望队列长度,而且减少了触发次数,节约了网络资源.最后,仿真验证和分析表明了所提方法的有效性和优越性.     

基于DRNet的常微分方程模型逼近和序列预测方法

针对残差网络预测精度偏低的问题,基于系统观测数据和相轨线的关系,提出密集残差网络的方法实现对自治系统的拟合逼近和序列的高精度预测。首先,为强化对数据内含“特征信息”的提取和流通使用,将神经网络各隐藏层的输入与之前各层输出拼接后作为本层的输入,形成密集连接模块;其次,为避免加大网络深度时出现的“退化”现象,引入残差机制,将密集连接模块的输入层与输出层相连,形成密集残差网络。最后,将密集残差网络应用于线性的单自由度系统振动模型和非线性的SEIRS模型、Logistic Volterra模型。结果表明,在规模为5 000和10 000的数据集上,密集残差网络对模型的拟合逼近效果和预测精度优于残差网络、反向传播神经网络和密集网络,特别是在非线性系统上的4项定量评价指标均优于对照模型,表现出密集残差网络对自治系统模型逼近和序列预测的高有效性;同时,在观测数据中加入5%的噪声后,密集残差网络表现出较好的抗干扰性。     

A Finite Element Algorithm for Nematic Liquid Crystal Flow Based on the Gauge-Uzawa Method

In this paper,we present a finite element algorithm for the time-dependent nematic liquid crystal flow based on the Gauge-Uzawa method.This algorithm combines the Gauge and Uzawa methods within a finite element variational formulation,which is a fully discrete projection type algorithm,whereas many projection methods have been studied without space discretization.Besides,error estimates for velocity and molecular orientation of the nematic liquid crystal flow are shown.Finally,numerical results are given to show that the presented algorithm is reliable and confirm the theoretical analysis.     

Random-phase approximation correlation methods for molecules and solids

Random-phase approximation (RPA) correlation methods based on Kohn–Sham density-functional theory and Hartree–Fock are derived using the adiabatic-connection fluctuation dissipation theorem. It is shown that the correlation energy within the adiabatic-connection fluctuation-dissipation theorem is exact in a Kohn–Sham framework while for Hartree–Fock reference states this is not the case. This shows that Kohn–Sham reference states are probably better suited to describe electron correlation for use in RPA methods than Hartree–Fock reference states. Both, Kohn–Sham and Hartree–Fock RPA methods are related to each other both by comparing the underlying correlation functionals and numerically through the comparison of total energies and reaction energies for a set of small organic molecules.     

An accurate theory for the square-well potential: from long to short well width

Density function theory (DFT) is combined with the first-order mean spherical approximation (FMSA) to study the radial distribution function (RDF) of the square-well (SW) potential. The combination (DFT + FMSA) is based on the direct correlation function (DCF) of the FMSA. Upon comparison with computer simulation data, DFT + FMSA is shown to give better performance than FMSA for mid- and long-range attractions. For short-range and very short-range attractions, the theory successfully corrects the deficiencies of the original FMSA. Comparisons include the evaluation of contact values, gap height at a discontinuity and profiles of the RDF. This work provides an accurate and consistent way to handle the SW potential.