对称正定Toeplitz方程组的多级迭代求解

二级迭代法亦称内外迭代法. 多级迭代法由多个二级迭代嵌套而成.这些方法特别适合于并行计算,同时可以理解为古典迭代法的延伸或共轭梯度法的预处理子.本文讨论了对称正定Toeplitz线性方程组多级迭代法. 首先,基于Toeplitz矩阵的结构, 我们给出了多级块Jacobi分裂,然后证明了每一级分裂均为P-正则分裂, 并证明了当每一级内迭代次数均为偶数时,迭代法的收敛性. 最后通过数值实例验证了此方法的有效性.     

实对称正定Toeplitz矩阵的带位移的Sine预处理子

本文研究了求解实对称正定Toeplitz线性方程组的预处理共轭梯度法.基于实对称Toeplitz矩阵都有一个三角变换分裂(TTS)的事实,我们提出了带位移的Sine预处理子TS,分析了预处理矩阵的谱性质,并讨论了每步迭代的计算复杂度.数值实验表明该预处理子比T.Chan预处理子~([2])更有效.     

关于二阶椭圆方程区域分裂法的最优预处理

§0.引言 区域分裂是与微分方程数值解的并行计算的数学基础密切相关的,预处理共轭梯度法是区域分裂的一个主要途径,寻找好的预处理子是关键问题,本文给出一个较一般性的方法,预处理过程包括一个整体小规模问题和若干个独立的局部子问题,整体问题和局部问题的选取均有极大的任意性,预处理条件数的估计是由整体问题和局部问题的一些特     

几乎各向同性的高维空间分数阶扩散方程的分块快速正则Hermite分裂预处理方法

一类空间分数阶扩散方程经过有限差分离散后所得到的离散线性方程组的系数矩阵是两个对角矩阵与Toeplitz型矩阵的乘积之和.在本文中,对于几乎各向同性的二维或三维空间分数阶扩散方程的离散线性方程组,采用预处理Krylov子空间迭代方法,我们利用其系数矩阵的特殊结构和具体性质构造了一类分块快速正则Hermite分裂预处理子.通过理论分析,我们证明了所对应的预处理矩阵的特征值大部分都聚集于1的附近.数值实验也表明,这类分块快速正则Hermite分裂预处理子可以明显地加快广义极小残量(GMRES)方法和稳定化的双共轭梯度(BiCGSTAB)方法等Krylov子空间迭代方法的收敛速度.     

速度追踪问题中鞍点系统的新分裂迭代

 有限元离散一类速度追踪问题后得到具有鞍点结构的线性系统,针对该鞍点系统,本文提出了一种新的分裂迭代技术.证明了新的分裂迭代方法的无条件收敛性,详细分析了新的分裂预条件子对应的预处理矩阵的谱性质.数值结果验证了对于大范围的网格参数和正则参数,新的分裂预条件子在求解有限元离散速度追踪问题得到的鞍点系统时的可行性和有效性.     

速度追踪问题产生的鞍点系统的新的分裂迭代技术

有限元离散一类速度追踪问题后得到具有鞍点结构的线性系统,针对该鞍点系统,本文提出了一种新的分裂迭代技术.证明了新的分裂迭代方法的无条件收敛性,详细分析了新的分裂预条件子对应的预处理矩阵的谱性质.数值结果验证了对于大范围的网格参数和正则参数,新的分裂预条件子在求解有限元离散速度追踪问题得到的鞍点系统时的可行性和有效性.     

关于H-矩阵的H-预处理子（英文）

设A为一实对称正定的严格对角占优矩阵.设A=D-B为A的Jacobi分裂.为了求解线性方程组Ax=b,在新提出的预处理子的基础上,我们采用预处理共轭梯度方法(PCG)来求解该问题.新提出的预处理子Pv=D+νvv~T,其中v=|B|e,e=(1,...,1)~T,ν=v~TBv/||v||_2~4,且ν使||cvv~T-B||_F达到极小.我们得到了预处理矩阵P_v~(-1)A特征值的上下界,它的界比JIN提出的预处理子的界简单紧凑.数值结果表明我们的预处理子的有效性.     

非埃米特正定Toeplitz矩阵的m—步预处理子（英文）

众所周知,如果A是Toeplitz矩阵,那么矩阵A有一循环与反循环分裂(记为CSCS)[7],可写为A=C+S,其中C为循环矩阵,S为反循环矩阵.本文针对某类Toeplitz矩阵,提出了一个m步的预处理子P_m,这个预处理子P_m是基于CSCS迭代方法构建的.本文中证明当C和S都是正定矩阵时,对于适当的m,预处理矩阵(P_m*A)**(P_m*A)的谱半径聚集于1.实验结果表明,对于适当的m,本文提出的预处理子优于T—Chan预处理子[3].     

广义鞍点问题的松弛维数分解预条件子

本文将Benzi等提出的松弛维数分解(Relaxed dimensionalfactorization, RDF)预条件子进一步推广到广义鞍点问题上,并称为GRDF(Generalized RDF)预条件子.该预条件子可看做是用维数分裂迭代法求解广义鞍点问题而导出的改进维数分裂(Modified dimensional split, MDS)预条件子的松弛形式, 它相比MDS预条件子更接近于系数矩阵, 因而结合Krylov子空间方法(如GMRES)有更快的收敛速度.文中分析了GRDF预处理矩阵特征值的一些性质,并用数值算例验证了新预条件子的有效性.     

对称正定矩阵的多级迭代法

本文主要研究解对称正定矩阵的多级迭代法,并对其收敛性进行证明。最后用数值实验验证此方法的有效性．多级迭代法特别适用于并行计算,并且可以被理解为古典迭代法的扩展,或共轭梯度法的预处理子。     

Wolfe线搜索下一类混合共轭梯度法的全局收敛性

本文给出了一个新的共轭梯度公式,新公式在精确线搜索下与DY公式等价,并给出了新公式的相关性质.结合新公式和DY公式提出了一个新的混合共轭梯度法,新算法在Wolfe线搜索下产生一个下降方向,并证明了算法的全局收敛性,并给出了数值例子.     

Exponentially-fitted Runge-Kutta-Nyström method for the numerical solution of initial-value problems with oscillating solutions

A new approach for constructing efficient Runge-Kutta-Nyström methods is introduced in this paper. Based on this new approach a new exponentially-fitted Runge-Kutta-Nyström fourth-algebraic-order method is obtained for the numerical solution of initial-value problems with oscillating solutions. The new method has an extended interval of periodicity. Numerical illustrations on well-known initial-value problems with oscillating solutions indicate that the new method is more efficient than other ones.     

一类由概率分布生成的新度量

主要得到了一类由概率分布生成的新度量.以信息理论中的重要概念-相对熵为基础,对前人文章中的重要结论进行推广,通过利用改进的初等方法在离散的可测空间中得到了这类新度量,并且证明了得到的新度量成立的充要条件.由此再将这类新度量推广到连续的可测空间中,得到了同样的结果.最后讨论了新度量的最值问题.     

Understanding the impacts of product knowledge and product type on the accuracy of intentions-based new product predictions

New product development involves several critical decisions. A key decision making area in new product development is the evaluation of the viability and the market potentials of a new product. In the absence of any relevant historical data, companies ask the potential buyers of their products about their intentions to buy those products when assessing their viability. Despite the popularity of the use of behavioral intentions in predicting the market acceptance of new product ideas, both survey and empirical studies suggest that the accuracy of such predictions is usually very low. Although earlier case-based studies suggest that a number of factors can affect the quality of new product decisions, it is still empirically unclear how product knowledge and the type of new products might impact the predictive accuracy of intentions-based new product forecasting. This study utilized a longitudinal research design and empirically tested the hypotheses across two new products. The study first collected purchase intentions data about the new products. Second, it collected subsequent actual purchase data about the new products. The results of series of hierarchical regression analyses comparing the initial purchase intentions and subsequent actual behaviors showed that while product knowledge is positively related to the predictive accuracy and consistency of intentions-based new product forecasting, product type is negatively related to them.     

一种非单调滤子信赖域算法解线性不等式约束优化

本文给出了一种新的多维滤子算法结合非单调信赖域策略解线性约束优化.目标函数及其投影梯度的分量组成了新的多维滤子，并且与信赖域半径有关.当信赖域半径充分小时，新的滤子能接受试探点，避免算法无限循环.非单调信赖域策略保证了新算法的整体收敛性.目前为止，多维滤子算法局部收敛性分析仍然没有解决，在合理假设下，我们分析了新算法的局部超线性收敛性.数值结果验证了算法的有效性.     

A gradient-related algorithm with inexact line searches

In this paper, a new gradient-related algorithm for solving large-scale unconstrained optimization problems is proposed. The new algorithm is a kind of line search method. The basic idea is to choose a combination of the current gradient and some previous search directions as a new search direction and to find a step-size by using various inexact line searches. Using more information at the current iterative step may improve the performance of the algorithm. This motivates us to find some new gradient algorithms which may be more effective than standard conjugate gradient methods. Uniformly gradient-related conception is useful and it can be used to analyze global convergence of the new algorithm. The global convergence and linear convergence rate of the new algorithm are investigated under diverse weak conditions. Numerical experiments show that the new algorithm seems to converge more stably and is superior to other similar methods in many situations.     

GLOBAL CONVERGENCE RESULTS OF A THREE TERM MEMORY GRADIENT METHOD WITH A NON-MONOTONE LINE SEARCH TECHNIQUE

In this paper, a new class of three term memory gradient method with nonmonotone line search technique for unconstrained optimization is presented. Global convergence properties of the new methods are discussed. Combining the quasi-Newton method with the new method, the former is modified to have global convergence property. Numerical results show that the new algorithm is efficient.     

Analysis of a new parareal algorithm based on waveform relaxation method for time-periodic problems

We present a new parallel algorithm for time-periodic problems by combining the waveform relaxation method and the parareal algorithm, which performs the parallelism both in sub-systems and in time. In the new algorithm, the waveform relaxation propagator is chosen as a new fine propagator instead of the classical fine propagator. And because of the characteristic of time-periodic problems, the new parareal waveform relaxation algorithm needs to solve a periodic coarse problem at the coarse level in each iteration. The new algorithm is proved to converge linearly at most. Then the theoretic parallel efficiency of the new algorithm is also considered. Numerical experiments confirm our analysis finally.     

A robust model for a leader–follower competitive facility location problem in a discrete space

This paper aims at determining the optimal locations for the leader’s new facilities under the condition that the number of the follower’s new facilities is unknown for the leader. The leader and the follower have some facilities in advance. The first competitor, the leader, opens p new facilities in order to increase her own market share. On the other hand, she knows that her competitor, the follower, will react to her action and locate his new facilities as well. The number of the follower’s new facilities is unknown for the leader but it is assumed that the leader knows the probability of opening different numbers of the follower’s new facilities. The leader aims at maximizing her own market share after the follower’s new facilities entry. The follower’s objective is also to maximize his own market share. Since the number of the follower’s new facilities is unknown for leader, “Robust Optimization” is used for maximizing the leader’s market share and making the obtained results “robust” in various scenarios in terms of different numbers of the follower’s new facilities. The optimal locations for new facilities of both the leader and the follower are chosen among pre-determined potential locations. It is assumed that the demand is inelastic. The customers probabilistically meet their demands from all different facilities and the demand level which is met by each facility is computed by Huff rule. The computational experiments have been applied to evaluate the efficiency of the proposed model.     

A new technique of integral representations in Cn

A new technique of integral representations in Cn, which is different from the well-known Henkin technique, is given. By means of this new technique, a new integral formula for smooth functions and a new integral representation of solutions of the -equations on strictly pseudoconvex domains in Cn are obtained. These new formulas are simpler than the classical ones, especially the solutions of the -equations admit simple uniform estimates. Moreover, this new technique can be further applied to arbitrary bounded domains in Cn so that all corresponding formulas are simplified.