一种基于积分方程的数值微分方法

本文研究了目前一些求解数值微分的方法无法求出端点导数或是求出的端点附近导数不可用的问题.利用构造一类积分方程的方法,将数值微分问题转化为这类积分方程的求解,并用一种加速的迭代正则化方法来求解积分方程. 数值实验结果表明该算法可以有效求出端点的导数,且具有数值稳定、计算简单等优点.     

求解脉冲电磁场的时序法

本文提出一种新的方法——时序展开法,来求解脉冲电磁场的激励或散射问题。以偶极子为例,按时间顺序,含电流的积分方程可以被分解为一系列递推积分方程。由于系列中各子方程间有递推关系,每一子方程的解具有形式简单的准行波特征;整个系列方程组便于依次求解. 为了求解各子方程,首先求无穷长单振子的电流响应;利用源函数的行波特性和电流沿导线传播的同时性,电流积分方程进一步分解成振幅、主波和长尾三个方程。它们可以独立依次求解,解的精度很高,仅一阶近似就达1％。     

浅谈二阶变系数齐次线性微分方程的解法

给出了二阶变系数齐次线性微分方程为恰当方程的充分与必要条件，对于恰当方程，给出了方程的求解方法.当二阶齐次线性方程不是恰当方程时，我们讨论了特殊情况下，如何求积分因子，进而把原来的方程变为恰当方程进行求解的方法.     

一类复合材料焊接的界面裂纹问题

利用复变方法和解析函数边值问题的基本理论，研究一类复合材料焊接线上出现裂纹的平面弹性基本问题，笔者通过适当的函数分解和积分变换，将寻找复应力函数的问题转化为求解一正而型奇异积分方程，并借助积分方程理论给出了方程的求解方法。     

求解第一类积分方程的正则化—小波方法及其数值试验

1 方法的描述 第一类(Fredholm)积分方程是指形如 (1.1)的积分方程,其中核k(x,y)和右端函数f(x)给定,u(x)是未知函数.许多物理、化学、力学和工程应用问题都能导致第一类积分方程.求解第一类积分方程的一个本质性困难是方程的不适定性,即解的存在性、唯一性和稳定性遭到破坏.常用的数值方法有奇异值分解(SVD)方法、Tikhonov正则化方法、投影方法、正则化-样条方法、再生核方法等.本文提出一种新的正则化-小波方法,在第一类积分方程有多个解时,可以求出具有最小范数的数值解;如果原积分方程有唯一解,则所得的数值解收敛于准确解.数值试验表明,该方法是可行的. 我们在L~2[a,b]中考虑第一类(Fredholm)积分方程,即假设方程(1.1)中积分算子K∈L~2([a,b]×[a,b])及右端f(x)∈L~2[a,b]给定.为保证数值求解算法的稳定性,我们先用正则化方法处理该方程,将不适定问题化为泛函极值问题来求解,然后利用多重正交样条小波基构造求解格式.由于我们给出了直接计算低阶的多重正交样条小波基函数的一般公式,使得解法可以在计算机迅速实现.     

带裂纹的弹性狭长体基本问题

利用复变方法和积分方程理论，讨论带任意裂纹的各向同性弹性狭长体的基本问题。通过适当的函数分解和积分变换，将问题简化为一正则型奇异积分方程。对方程解的情况和求解方法进行了研究，并导出裂纹尖端的应力强度因子。     

偏微分方程△^2u—s△u＋k^2u=0边值问题的MRM边界变分方程

导出边值问题△^2u-s△u k^2u=0；x∈Ω∪Ω‘∪→R^2;u/Γ=u0;δu/δn/Γ=g0的定解问题，MRM边界变分方程，全平面解的表达式，从中可以以看出，MRM边界变分方程中只包含弱奇异积分核，并且自动消除了原第一，二MRM边界积分方程中出现的强奇异积分核，问题解的表达式后并不加任何多项式，因而也不需要引入Lagrange乘子求解该项，这给边界元数值求解过程带来极大的方便，数值分析结果表明该方法具有明显优势。     

声学外问题的边界积分方程方法

本文考虑了二维Ｈｅｌｍｈｏｌｔｚ方程的Ｄｉｒｉｃｈｌｅｔ问题．首先利用单层位势将问题归结为第一类边界积分方程，然后利用有限元方法求解该边界积分方程的近似解，给出近似解的收敛性、一致收敛性以及最优收敛速度定理，并进行解的稳定性分析．     

三维瞬态弹性动力场的基本解及边界积分方程

三维瞬态动力场的求解,一直为世人所瞩目,本文拟采用边界积分方程——边界单元法进行求解。与其它数值方法相比较,本方法由于采用了动力场区域的基本解,因此在处理无限域或半无限域的动力场问题时可以毋须人为地去划定边界,附加相应的边界约束条件。同时在求解过程中,采用了权函数与奇异格林函数,使得域积分可以转化为边界积     

求解一类具有Hilbert核的奇异积分方程的小波方法

1 引言 近年来,用小波方法数值求解积分方程越来越引起人们的注意.文献[1]提出的算法可将一类积分算子所对应的矩阵稀疏化,为小波方法快速求解积分方程开辟了一条新的道路.     

Numerical solution of Fredholm integral equations of the second kind by using integral mean value theorem II. High dimensional problems

In this work, we generalize the numerical method discussed in [Z. Avazzadeh, M. Heydari, G.B. Loghmani, Numerical solution of Fredholm integral equations of the second kind by using integral mean value theorem, Appl. math. modelling, 35 (2011) 2374–2383] for solving linear and nonlinear Fredholm integral and integro-differential equations of the second kind. The presented method can be used for solving integral equations in high dimensions. In this work, we describe the integral mean value method (IMVM) as the technical algorithm for solving high dimensional integral equations. The main idea in this method is applying the integral mean value theorem. However the mean value theorem is valid for multiple integrals, we apply one dimensional integral mean value theorem directly to fulfill required linearly independent equations. We solve some examples to investigate the applicability and simplicity of the method. The numerical results confirm that the method is efficient and simple.     

On the solution of a class of integral equations

In this paper we consider a class of Fredholm integral equations of the first kind which arise in a large number of problems in applied mathematics. Although only certain special cases of the equations can be solved exactly, it is shown that a constructive method can be developed for reformulating the equations as Fredholm integral equations of the second kind. This approach will be seen to cover and bring together the large number of isolated cases of the equations which have appeared in the literature. Several examples are given to illustrate the general method.     

Canonical Boundary Element Method for Plane Elasticity Problems

In this paper, we apply the canonical boundary reduction, suggested by Feng Kang, to the plane elasticity problems, find the expressions of canonical integral equations and Poisson integral formulas in some typical domains. We also give the numerical method for solving these equations together with their convergence and error estimates. Coupling with classical finite element method, this method can be applied to other domains.     

积分算子方程的连续型配置方法

1．引言由于对积分算子方程来说，配置法比Galerkin法具计算量小的优点（少算一重积分），故配置法更受人们重视．但已有的文献几乎都是将配置空间取作非连续的分片多项式样条空间，以得到某种超收敛结果（如［1，2］）．这种方法存在下列不足：（a）光滑核Volterra积分方程与光滑核Fredholm积分方程具完全不同的收敛性质［1］，且需用不同的方法获得其加速收敛结果（比较［31与［4］），尽管Volterra积分方程在理论上被看作是Fredholm积分方程的特殊情形；（b）光滑核Volterra积分方程的配置解不具任何超收敛性，其迭代配置解也只在结点…     

一维热传导方程移动边界识别的计算方法

构造了一种正则化的积分方程方法来由Cauchy数据确定一维热传导方程的移动边界．在将区域延拓至规则区域后,通过Fourier方法将问题转化为一个第一类Volterra积分方程．然后分别用Lavrentiev正则化方法以及Tikhonov正则化方法将不稳定的第一类Volterra积分方程转化为适定的第二类积分方程,并分别将积分方程转化为常微分方程组,并用Runge—Kutta方法数值求解,以及直接离散来求解．最后通过自由边界上的条件得到数值的移动边界．通过一些数值试验表明此方法是有效可行的,并且给出的方法无需迭代,数值计算较简单．     

Improving the accuracy of solutions of the linear second kind volterra integral equations system by using the taylor expansion method

This paper presents a new and an efficient method for determining solutions of the linear second kind Volterra integral equations system. In this method, the linear Volterra integral equations system using the Taylor series expansion of the unknown functions transformed to a linear system of ordinary differential equations. For determining boundary conditions we use a new method. This method is effective to approximate solutions of integral equations system with a smooth kernel, and a convolution kernel. An error analysis for the proposed method is provided. And illustrative examples are given to represent the efficiency and the accuracy of the proposed method.     

Optimal and Approximate Solutions of Singular Integral Equations by Means of Reproducing Kernels

A reproducing kernel method is proposed to obtain the optimal and approximate solutions of Carleman singular integral equations. Therefore, we will be mostly interested in singular integral equations with a Cauchy type kernel and whose coefficients are real or complex valued functions. The new method and corresponding concepts allow the analysis of associated discrete singular integral equations and corresponding inverse source problems in appropriate frameworks.     

Volterra积分方程的蒙特卡罗数值求解方法

本文讨论了第一类、第二类以及具有奇异核的Volterra积分方程的数值解问题.利用重要抽样蒙特卡罗随机模拟方法获得积分方程解的近似计算结果.通过对文献中算例的实现表明文中所提方法扩展了Volterra型积分方程的数值求解方法,     

On nonlinear Fredholm integral equations with non-differentiable Nemystkii operator

From decomposition method for operators, we consider a Newton-Steffensen iterative scheme for approximating a solution of nonlinear Fredholm integral equations with non-differentiable Nemystkii operator. By means of a convergence study of the iterative scheme applied to this type of nonlinear Fredholm integral equations, we obtain domains of existence and uniqueness of solution for these equations. In addition, we illustrate this study with a numerical experiment.     

A boundary element method to price time-dependent double barrier options

In this paper we propose a new method for pricing double-barrier options with moving barriers under the Black-Scholes and the CEV models. First of all, by applying a variational technique typical of the boundary element method, we derive an integral representation of the double-barrier option price in which two of the integrand functions are not given explicitly but must be obtained solving a system of Volterra integral equations of the first kind. Second, we develop an ad hoc numerical method to regularize and solve the system of integral equations obtained. Several numerical experiments are carried out showing that the overall algorithm is extraordinarily fast and accurate, even if the barriers are not differentiable functions. Moreover the numerical method presented in this paper performs significantly better than the finite difference approach.