广西GDP的统计预测模型及其应用

本文利用SPSS统计软件及非参数统计方法(卡方检验和K-S检验法)对广西1950年至2006年共57年的GDP数据进行实证分析.在最佳准则(即AIC准则)下建立了ARIMA(1,2,1)时间序列模型,并利用非参数统计方法对此模型进行了适应性检验,然后利用2001年至2006年的实际值与该模型的预测值进行了比较.最后,本文利用该模型对广西未来五年的GDP进行了预测.     

非均匀随机数产生

本文详细介绍了产生非均匀随机数的一般方法,常用连续分布(正态分布,Gamma分布,Beta分布,χ2分布和F分布)的抽样法及利用R软件产生随机数的方法。     

一种新的多目标人工蜂群算法

提出了一种新的多目标人工蜂群算法(GMOABC),该算法使用自适应网格维护外部档案.在雇佣蜂阶段,选择外部档案中的个体作为食物源的"邻居",用于生成新食物源.在观察蜂阶段,除了"邻居",全局最优食物源gbest也用于指导观察蜂的飞行轨迹.借助Friedman非参数检验法将GMOABC算法与其他主流算法在CEC 2009测试函数集上进行了全面比较,实验结果显示了文章算法的优越性.     

一种求解大型Lyapunov矩阵方程的预处理并行算法

研究了一种求解大型Lyapunov矩阵方程的并行预处理变形共轭梯度法.首先将处理小型矩阵方程的Smith预处理方法引入该问题的求解,将原矩阵方程转变为Stein方程,然后采用变形共轭梯度法并行求解预处理后的矩阵方程.其中遇到的难点是需要确定参数μ及求矩阵(A+μI)的逆.基于估计特征值的Gerschgorin圆定理给出了参数μ的估值,再采用变形共轭梯度法并行求得矩阵(A +μ l)的逆,从而形成预处理后的矩阵方程.通过数值试验,该算法与未预处理的变形共轭梯度法相比较,预处理算法明显优于未预处理的算法,而且其并行效率高达0.85.     

垂直线性互补问题的一种光滑算法

文中给出了垂直线性互补问题的一个新的光滑价值函数,不同于光滑化方法中的价值函数,它不包含任何必须趋向零的参数,因此算法中不涉及参数调整步骤,而且具有良好的强制性.基此价值函数,提出了求解垂直线性互补问题的一种阻尼Newton类算法,并证明了该算法对竖块P0+R0矩阵的垂直线性互补问题具有全局收敛性;当解满足相当于BD-正则条件时,算法具有局部二次收敛性;在不增加额外校正步骤(算法的每个迭代步只求解一个Newton方程)的情形下,算法对竖块P-矩阵垂直线性互补问题(无须假设严格互补),具有有限步收敛性.数值实验结果令人满意.     

基于Anderson-Darling正态性检验的改进方法及应用

首先介绍了Anderson-Darling正态性检验法的基本原理,并应用稳健统计方法——用中位值(x_med)估计总体均值μ、用标准四分位间距(NIQR)估计总体标准偏差σ,来改进Anderson-Darling统计量(AD统计量)以增强其稳健性.其次利用蒙特卡罗(Monte-Carlo)随机模拟方法产生正态分布随机数,计算改进的AD统计量,重新确定改进的AD统计量的上α分位数以及对应的上尾显著点,进而给出上尾显著点表格和拟合函数曲线供统计人员使用.以长春黄金研究院有限公司在2021年度开展的金锭中金含量化学分析能力验证为例,给出了改进后该统计量的具体应用过程,结果令人满意.     

偏微分方程(组)完全对称分类微分特征列集算法

给出了一个确定含参数偏微分方程(组)的完全对称分类微分特征列集算法,该算法能够直接、系统地确定偏微分方程(组)的完全对称分类.用给出的算法获得了含任意函数类参数的线性和非线性波动方程完全势对称分类.这也是微分形式特征列集算法(微分形式吴方法)在微分方程领域中的新应用.     

Marshall-Olkin威布尔分布的贝叶斯分析（英文）

Kundu与Gupta~([1])提出用重要抽样法来计算Marshal-Olkin两元威布尔分布参数的贝叶斯估计,然而我们发现在样本量变大的情况下,重要抽样算法的参数估计效果却不理想.在这篇文章中,我们引入潜在变量来简化似然函数,并且提出利用MCMC算法实现对该模型未知参数的估计.为了评价我们提出方法的有效性,我们将提出的贝叶斯方法与极大似然估计数据模拟结果作对比,可以发现:即使在样本量很小的情况下,提出的贝叶斯方法的参数估计效果更理想.实例分析也说明了这一点.     

(3+1)维变系数Kudryashov-Sinelshchikov(K-S)方程的同宿呼吸波解和高阶怪波解

基于Hirota双线性方法,利用拓展的同宿呼吸检验法得到了(3+1)维变系数Kudryashov-Sinelshchikov(K-S)方程的同宿呼吸波解,对该解的参数选取合适的数值,可得到不同结构的同宿呼吸波.通过对同宿呼吸波解的周期取极限,推导出方程的怪波解.最后,构造出一个特殊的高阶多项式作为测试函数,求得该方程的一阶怪波解和二阶怪波解.     

一个充分下降的改进HS共轭梯度法

共轭梯度法是求解大规模无约束优化问题最有效的方法之一.对HS共轭梯度法参数公式进行改进,得到了一个新公式,并以新公式建立一个算法框架.在不依赖于任何线搜索条件下,证明了由算法框架产生的迭代方向均满足充分下降条件,且在标准Wolfe线搜索条件下证明了算法的全局收敛性.最后,对新算法进行数值测试,结果表明所改进的方法是有效的.     

The generalized Gompertz distribution

This paper deals with a new generalization of the exponential, Gompertz, and generalized exponential distributions. This distribution is called the generalized Gompertz distribution (GGD). The main advantage of this new distribution is that it has increasing or constant or decreasing or bathtub curve failure rate depending upon the shape parameter. This property makes GGD is very useful in survival analysis. Some statistical properties such as moments, mode, and quantiles are derived. The failure rate function is also derived. The maximum likelihood estimators of the parameters are derived using a simulations study. Real data set is used to determine whether the GGD is better than other well-known distributions in modeling lifetime data or not.     

一个求解线性规划的单纯形-内点算法

根据单纯形方法和大步长路径跟踪算法（Ｈｅｒｔｏｇ，Ｒｏｏｓ和Ｔｅｒｌａｋｙ１９９１），对于具有不等式约束的线性规划问题，引进了一个具有组合特性的内点算法．该方法保留了单纯形方法和内点算法的优点，克服了它们的不足，在任何情况下，这个方法都能快速收敛．数值结果也很好地验证了这个结论．     

A numerical approach for a semilinear parabolic equation with a nonlocal boundary condition

A semilinear reaction-diffusion problem with a nonlocal boundary condition is studied. This paper presents a new and very easy implementable numerical algorithm for computations. This is based on a suitable linearization in time and on the principle of linear superposition. Any method for the space discretization (FEM was taken in this analysis) can be chosen. The derived algorithm is implicit and it does not need any iteration scheme to get a solution with the nonlocal boundary condition. Stability analysis has been performed and the optimal error estimates have been derived. Numerical results have been compared with other known techniques.     

Thresholds of the Divisor Methods

We give an algorithm which permits calculating the maximum and minimum vote shares that allow a party to obtain h seats, that is, the threshold of exclusion and the threshold of representation. These have already been studied for some methods (such as d'Hondt or Sainte-Laguë), and are here generalized to any divisor method, and to any number of seats. The thresholds depend on the size of the constituency, the number of parties running in the constituency, and the divisor method used. Finally, we give some consequences, including a characterization of the d'Hondt method.     

系统可靠性分配的综合因子法

综合因子法是系统可靠性分配的一种基本方法 ,本文在现有方法综合因子法的基础上概括出其可靠性分配的基本思想 ,使该方法可对单调关联系统进行可靠性分配 ,在计算机上予以实现 ,并在文中举例加以说明     

The method of fundamental solutions with eigenfunction expansion method for nonhomogeneous diffusion equation

In this article we describe a numerical method to solve a nonhomogeneous diffusion equation with arbitrary geometry by combining the method of fundamental solutions (MFS), the method of particular solutions (MPS), and the eigenfunction expansion method (EEM). This forms a meshless numerical scheme of the MFS‐MPS‐EEM model to solve nonhomogeneous diffusion equations with time‐independent source terms and boundary conditions for any time and any shape. Nonhomogeneous diffusion equation with complex domain can be separated into a Poisson equation and a homogeneous diffusion equation using this model. The Poisson equation is solved by the MFS‐MPS model, in which the compactly supported radial basis functions are adopted for the MPS. On the other hand, utilizing the EEM the diffusion equation is first translated to a Helmholtz equation, which is then solved by the MFS together with the technique of the singular value decomposition (SVD). Since the present meshless method does not need mesh generation, nodal connectivity, or numerical integration, the computational effort and memory storage required are minimal as compared with other numerical schemes. Test results for two 2D diffusion problems show good comparability with the analytical solutions. The proposed algorithm is then extended to solve a problem with irregular domain and the results compare very well with solutions of a finite element scheme. Therefore, the present scheme has been proved to be very promising as a meshfree numerical method to solve nonhomogeneous diffusion equations with time‐independent source terms of any time frame, and for any arbitrary geometry. © 2006 Wiley Periodicals, Inc. Numer Methods Partial Differential Eq, 2006     

The numerical solution of linear time-dependent partial differential equations by the Laplace transform and finite difference approximations

A feasible method is presented for the numerical solution of a large class of linear partial differential equations which may have source terms and boundary conditions which are time-varying. The Laplace transform is used to eliminate the time-dependency and to produce a subsidiary equation which is then solved in complex arithmetic by finite difference methods. An effective numerical Laplace transform inversion algorithm gives the final solution at each spatial mesh point for any specified set of values of t. The single-step property of the method obviates the need to evaluate the solution at a large number of unwanted intermediate time points. The method has been successfully applied to a variety of test problems and, with two alternative numerical Laplace transform inversion algorithms, has been found to give results of good to excellent accuracy. It is as accurate as other established finite difference methods using the same spatial grid. The algorithm is easily programmed and the same program handles equations of parabolic and hyperbolic type.     

Some investigations in a new algorithm for nonlinear optimization based on conic models of the objective function

An appealing approach to the solution of nonlinear optimization problems based on conic models of the objective function has been recently introduced by Davidon. It leads to a broad class of algorithms which, in some sense, can be considered to generalize the existing quasi-Newton algorithms. One particular member of this class has been deeply examined by Sorensen, who has proved some interesting theoretical properties. A new interpretation of this algorithm is suggested in this paper from a more straightforward and somewhat familiar point of view. In addition, numerical experiments have been carried out to compare the Sorensen algorithm with a straightforward BFGS implementation of the classical quasi-Newton method with the final aim to assess the real merits and benefits of the new algorithm. Although some challenging test functions are used in computational experiments, the results are not particularly favorable to the new algorithm. As a matter of fact they do not exhibit any jump of quality, as it might be expected. Lastly, it is pointed out that a difficulty may affect the new method in situations in which it is necessary to exploit the special structure of large-scale problems.This work was supported by the National Research Council of Italy under Grant No. 80-01144.     

Parallel Variational Bayes for Large Datasets With an Application to Generalized Linear Mixed Models

The article develops a hybrid variational Bayes (VB) algorithm that combines the mean-field and stochastic linear regression fixed-form VB methods. The new estimation algorithm can be used to approximate any posterior without relying on conjugate priors. We propose a divide and recombine strategy for the analysis of large datasets, which partitions a large dataset into smaller subsets and then combines the variational distributions that have been learned in parallel on each separate subset using the hybrid VB algorithm. We also describe an efficient model selection strategy using cross-validation, which is straightforward to implement as a by-product of the parallel run. The proposed method is applied to fitting generalized linear mixed models. The computational efficiency of the parallel and hybrid VB algorithm is demonstrated on several simulated and real datasets. Supplementary material for this article is available online.     

A new genetic approach to construct near-optimal binary search trees

Many definitive and approximate methods have been so far proposed for the construction of an optimal binary search tree. One such method is the use of evolutionary algorithms with satisfactorily improved cost efficiencies. This paper will propose a new genetic algorithm for making a near-optimal binary search tree. In this algorithm, a new greedy method is used for the crossover of chromosomes while a new way is also developed for inducing mutation in them. Practical results show a rapid and desirable convergence towards the near-optimal solution. The use of a heuristic to create not so costly chromosomes as the first offspring, the greediness of the crossover, and the application of elitism in the selection of future generation chromosomes are the most important factors leading to near-optimal solutions by the algorithm at desirably high speeds. Due to the practical results, increasing problem size does not cause any considerable difference between the solution obtained from the algorithm and exact solution.