弦割法与Muller法收敛阶的新证明方法

弦割法、Muller法与牛顿法一样,都是求解非线性方程的著名算法之一.然而在目前众多优秀的数值分析教材或论著中.关于弦割法和Muller法收敛阶的证明过程都是比较复杂的,无一例外的都是借助于差分方程的求解.本文对这两个算法的收敛阶给出了一种新的简单、直接的证明方法,达到了与牛顿法收敛阶证明方法的统一,同时还能够方便地求...     

A new application of Lagrange-Bürmann expansions. I. General principle

A new scheme is developed for improving the convergence of slowly convergent series solutions. The method is based on a transformation of variables of similarity form whereby the resulting composite function is constructed by its Lagrange-Bürmann expansion. It is the improved convergence of the new expansion that we take most advantage of in this method. The convergence of the Lagrange-Bürmann expansion as well as its inversion scheme is proved for analytic (object) functions. The inversion is required to recover from the Lagrange-Bürmann expansion the object function which is imbedded in the mapping functions. Several numerical examples demonstrate the improved convergence of the new method. The improvement owes much to the invariance properties of the mapping function under a group and the “built-in” feature of analytic continuation of the method. These features are elucidated in detail.     

Analysis of the quasi-Laguerre method

The quasi-Laguerre's iteration formula, using first order logarithmic derivatives at two points, is derived for finding roots of polynomials. Three different derivations are presented, each revealing some different properties of the method. For polynomials with only real roots, the method is shown to be optimal, and the global and monotone convergence, as well as the non-overshooting property, of the method is justified. Different ways of forming quasi-Laguerre's iteration sequence are addressed. Local convergence of the method is proved for general polynomials that may have complex roots and the order of convergence is . Received June 30, 1996 / Revised version received August 12, 1996     

NONCONFORMING FINITE ELEMENT PENALTY METHOD FOR STOKES EQUATION

a special penalty method is presented to improve the accuracy of the standard penaltymethod (or solving Stokes equation with nonconforming finite element, It is shown that thismethod with a larger penalty parameter can achieve the same accuracy as the staodaxd methodwith a smaller penalty parameter. The convergence rate of the standard method is just hall order of this penalty method when using the same penalty parameter, while the extrapolationmethod proposed by Faik et al can not yield so high accuracy of convergence. At last, we alsoget the super-convergence estimates for total flux.     

Conformal mapping by the method of alternating projections

Summary The functional analytic principle of alternating projections is used to construct an iterative method for numerical conformal mapping of the unit disc onto regions with smooth boundaries. The result is a simple method which requires in each iterative step only two complex Fourier transforms. Local convergence can be proved using a theorem of Ostrowski. Convergence is linear. The asymptotic convergence factor is equal to the spectral radius of a certain operator. A version with overrelaxation as well as a discretized version are discussed along the same lines. For regions which are close to the unit disc convergence is fast. For some familiar regions the convergence factors can be calculated explicitly. Finally, the method is compared with Theodorsen's.Dedicated to the memory of Peter Henrici     

从常步长梯度方法的视角看不可微凸优化增广Lagrange方法的收敛性

增广Lagrange方法是求解非线性规划的一种有效方法.从一新的角度证明不等式约束非线性非光滑凸优化问题的增广Lagrange方法的收敛性.用常步长梯度法的收敛性定理证明基于增广Lagrange函数的对偶问题的常步长梯度方法的收敛性,由此得到增广Lagrange方法乘子迭代的全局收敛性.     

On superlinear convergence of quasi-Newton methods for nonsmooth equations

We show that strong differentiability at solutions is not necessary for superlinear convergence of quasi-Newton methods for solving nonsmooth equations. We improve the superlinear convergence result of Ip and Kyparisis for general quasi-Newton methods as well as the Broyden method. For a special example, the Newton method is divergent but the Broyden method is superlinearly convergent.     

一种改进灰狼优化算法研究及应用

针对灰狼算法易陷入局部最优、收敛精度不高、收敛速度慢等缺点,提出一种改进的灰狼算法.引入莱维飞行,扩大搜索范围,增强全局搜索能力,避免陷入局部最优;引入贪婪原理,提升种群优良性以提高算法收敛精度;引入自适应收敛因子,加快收敛速度;引入动态权重策略,制约全局搜索与局部搜索的相互影响.将改进算法与其他四种算法作对比,实验表...     

Inexact-Newton method for solving operator equations in infinite-dimensional spaces

In this paper, we develop an inexact-Newton method for solving nonsmooth operator equations in infinite-dimensional spaces. The linear convergence and superlinear convergence of inexact-Newton method under some conditions are shown. Then, we characterize the order of convergence in terms of the rate of convergence of the relative residuals. The present inexact-Newton method could be viewed as the extensions of previous ones with same convergent results in finite-dimensional spaces.     

A Truncated Nonmonotone Gauss-Newton Method for Large-Scale Nonlinear Least-Squares Problems

In this paper, a Gauss-Newton method is proposed for the solution of large-scale nonlinear least-squares problems, by introducing a truncation strategy in the method presented in [9]. First, sufficient conditions are established for ensuring the convergence of an iterative method employing a truncation scheme for computing the search direction, as approximate solution of a Gauss-Newton type equation. Then, a specific truncated Gauss-Newton algorithm is described, whose global convergence is ensured under standard assumptions, together with the superlinear convergence rate in the zero-residual case. The results of a computational experimentation on a set of standard test problems are reported.     

一维凸函数牛顿法的全局收敛性及其应用

牛顿法是求解非线性方程F（x)＝0的一种经典方法。在一般假设条件下，牛顿法只具有局部收敛性。本文证明了一维凸函数牛顿法的全局收敛性，并且给出了它在全局优化积分水平集方法中的应用。     

Local convergence of a secant type method for solving least squares problems

Local convergence of a secant type iterative method for approximating a solution of nonlinear least squares problems is investigated in this paper. The radius of convergence is determined as well as usable error estimates. Numerical examples are also provided.     

A convergence analysis of the iterative aggregation method with one parameter

A method for accelerating linear iterations in a Banach space is studied as a linear iterative method in an augmented space, and sufficient conditions for convergence are derived in the general case and in ordered Banach spaces. An acceleration of convergence takes place if an auxiliary functional is chosen sufficiently close to a dual eigenvector associated with a dominant simple eigenvalue of the iteration operator; in this case, the influence of this eigenvalue on the asymptotic rate of convergence is eliminated. Quantitative estimates and bounds on convergence are given.     

Convergence of the Lavrent’ev method

The convergence of the Lavrent’ev method, which is a well-known regularization method for integral equations of the first kind, is analyzed as applied to equations with arbitrary linear bounded operators. A theorem concerning necessary and sufficient conditions for this convergence is proved. It is shown that these conditions are satisfied for two classes of integral equations that do not possess the properties required by the classical Lavrent’ev method.     

On the Convergence Analysis of a Two-Step Modification of the Gauss-Newton Method

We investigate the convergence of a two-step modification of the Gauss-Newton method applying the generalized Lipschitz condition for the first and second order derivatives. The convergence order as well as the convergence radius of the method are studied and the uniqueness ball of solution of the nonlinear least squares problem is examined. Finally, we carry out numerical experiments on a set of well-known test problems. (© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)     

On the quadratic convergence of the Aitken Δ<Superscript>2</Superscript> process

The Aitken Δ² method for finding fixed points of scalar mappings is interpreted as a modification of the Wegstein method. Based on this approach, conditions for the quadratic convergence of this method are obtained for various situations of convergence/divergence of simple iteration. An algorithm for calculating fixed points that keeps track of these situations is presented.     

A study on the convergence of homotopy analysis method

In this study, a reliable approach for convergence of the homotopy analysis method when applied to nonlinear problems is discussed. First, we present an alternative framework of the method which can be used simply and effectively to handle nonlinear problems. Then, mainly, we address the sufficient condition for convergence of the method. The convergence analysis is reliable enough to estimate the maximum absolute truncated error of the homotopy series solution. The analysis is illustrated by investigating the convergence results for some nonlinear differential equations. The study highlights the power of the method.     

m步非线性多重分裂Newton法的收敛性

本文讨论了多重分裂算法在求解一类非线性方程组的全局收敛性和单侧收敛性．当用研步Newton法来代替求得每个非线性多重分裂子问题的近似解时，同样给出相应收敛性结论．数值算例证实了算法的有效性．     

An alternating direction method of multipliers for elliptic equation constrained optimization problem

We propose an alternating direction method of multipliers (ADMM) for solving the state constrained optimization problems governed by elliptic equations. The unconstrained as well as box-constrained cases of the Dirichlet boundary control, Robin boundary control, and right-hand side control problems are considered here. These continuous optimization problems are transformed into discrete optimization problems by the finite element method discretization, then are solved by ADMM. The ADMM is an efficient first order algorithm with global convergence, which combines the decomposability of dual ascent with the superior convergence properties of the method of multipliers. We shall present exhaustive convergence analysis of ADMM for these different type optimization problems. The numerical experiments are performed to verify the efficiency of the method.     

Computing the region of convergence for power series in many real variables: A ratio-like test

We give an elementary proof that the region of convergence for a power series in many real variables is a star-convex domain but not, in general, a convex domain. In doing so, we deduce a natural higher-dimensional analog of the so-called ratio test from univariate power series. From the constructive proof of this result, we arrive at a method to approximate the region of convergence up to a desired accuracy. While most results in the literature are for rather specialized classes of multivariate power series, the method devised here is general. As far as applications are concerned, note that while theorems such as the Cauchy-Kowalevski theorem (and its generalizations to many variables) grant the existence of a region of convergence for a multivariate Taylor series to certain PDEs under appropriate restrictions, they do not give the actual region of convergence. The determination of the maximal region of convergence for such a series solution to a PDE is one application of our result.