Deconvolution and Regularization with Toeplitz Matrices

By deconvolution we mean the solution of a linear first-kind integral equation with a convolution-type kernel, i.e., a kernel that depends only on the difference between the two independent variables. Deconvolution problems are special cases of linear first-kind Fredholm integral equations, whose treatment requires the use of regularization methods. The corresponding computational problem takes the form of structured matrix problem with a Toeplitz or block Toeplitz coefficient matrix. The aim of this paper is to present a tutorial survey of numerical algorithms for the practical treatment of these discretized deconvolution problems, with emphasis on methods that take the special structure of the matrix into account. Wherever possible, analogies to classical DFT-based deconvolution problems are drawn. Among other things, we present direct methods for regularization with Toeplitz matrices, and we show how Toeplitz matrix–vector products are computed by means of FFT, being useful in iterative methods. We also introduce the Kronecker product and show how it is used in the discretization and solution of 2-D deconvolution problems whose variables separate.     

Confidence Intervals for Discrete Approximations to Ill-Posed Problems

We consider the linear model Y = Xβ + ε that is obtained by discretizing a system of first-kind integral equations describing a set of physical measurements. The n vector β represents the desired quantities, the m x n matrix X represents the instrument response functions, and the m vector Y contains the measurements actually obtained. These measurements are corrupted by random measuring errors ε drawn from a distribution with zero mean vector and known variance matrix. Solution of first-kind integral equations is an ill-posed problem, so the least squares solution for the above model is a highly unstable function of the measurements, and the classical confidence intervals for the solution are too wide to be useful. The solution can often be stabilized by imposing physically motivated nonnegativity constraints. In a previous article (O'Leary and Rust 1986) we developed a method for computing sets of nonnegatively constrained simultaneous confidence intervals. In this article we briefly review the simultaneous intervals and then show how to compute nonnegativity constrained one-at-a-time confidence intervals. The technique gives valid confidence intervals even for problems with m < n. We demonstrate the methods using both an overdetermined and an underdetermined problem obtained by discretizing an equation of Phillips (Phillips 1962).     

Integral equations of the first kind in the theory of oscillating plates

A modified matrix of fundamental solutions is used to derive and solve first-kind integral equations for the problem of high-frequency harmonic oscillations of an infinite elastic plate with a hole when Dirichlet or Neumann conditions are prescribed on the boundary curve.     

Reissner板中的夹杂和缺陷问题研究

基于保角变换技术和Faber级数展开,研究了含任意形状夹杂或缺陷的无限大Reissner板弯曲问题．将变换域中单位圆内、外解析函数分别展开成Faber级数,并将波动函数展开成第一类和第二类修正的n阶Bessel函数;利用边界位移、剪力和弯矩连续性条件得到问题的高阶线性方程组．以含椭圆形夹杂和缺陷的无限大Reissner板柱面弯曲为例,进一步给出了数值算例和理论分析．结果表明,对于软夹杂,板内力矩随夹杂与板厚尺寸比a/h变化非常敏感;在含硬夹杂条件下,板内力矩随夹杂尺寸变化相对不敏感．     

Polynomial Volterra integral equations of the first kind and the Lambert function

The role of the Lambert function in the theory of polynomial Volterra equations of the first kind is considered. New results are presented in addition to the known ones. In particular, the stability of a continuous solution of the first-kind polynomial Volterra equation of degree N is investigated. Based on the techniques of majorant equations, sufficient stability conditions are obtained.     

A Study of the stability of solutions to inverse problems of dynamics of control systems under perturbations of initial data

A novel dual approach to the problem of optimal correction of first-kind improper linear programming problems with respect to their right-hand sides is proposed. It is based on the extension of the traditional Lagrangian by introducing additional regularization and barrier components. Convergence theorems are given for methods based on the augmented Lagrangian, an informal interpretation of the obtained generalized solution is suggested, and results of numerical experiments are presented.     

New features of stability of linear systems of differential equations with constant coefficients

We consider stability of linear systems of differential equations with constant real coefficients whose matrices are off-diagonally non-negative. The results are applied to arbitrary linear systems of differential equations with constant complex coefficients.     

On the invariants of two dimensional linear parabolic equations

We consider the most general two dimensional linear parabolic equations. Motivated by the recent work of Ibragimov et al. [1], [2], [3] we construct differential invariants, semi-invariants and invariant equations. These results are achieved with the employment of the equivalence group admitted by this class of parabolic equations. We derive those variable coefficient equations of this class of linear parabolic equations that can be mapped into constant coefficient equations. Further applications are presented.     

Differential equations with 2-term recursion

In this paper we consider linear differential equations with a recursion consisting of two terms. We consider these equations in positive characteristics and in characteristic zero. We will find a new proof for the Grothendieck conjecture for these equations. Supported by a grant from the DFG.     

Further analysis on classifications of PDE(s) with variable coefficients

In this study we consider further analysis on the classification problem of linear second order partial differential equations with non-constant coefficients. The equations are produced by using convolution with odd or even functions. It is shown that the patent of classification of new equations is similar to the classification of the original equations.     

The trade-off between regularity and stability in Tikhonov regularization

When deriving rates of convergence for the approximations generated by the application of Tikhonov regularization to ill-posed operator equations, assumptions must be made about the nature of the stabilization (i.e., the choice of the seminorm in the Tikhonov regularization) and the regularity of the least squares solutions which one looks for. In fact, it is clear from works of Hegland, Engl and Neubauer and Natterer that, in terms of the rate of convergence, there is a trade-off between stabilization and regularity. It is this matter which is examined in this paper by means of the best-possible worst-error estimates. The results of this paper provide better estimates than those of Engl and Neubauer, and also include and extend the best possible rate derived by Natterer. The paper concludes with an application of these results to first-kind integral equations with smooth kernels.

     

Singularity Results for Functional Equations Driven by Linear Fractional Transformations

We consider functional equations driven by linear fractional transformations, which are special cases of de Rham’s functional equations. We consider Hausdorff dimension of the measure whose distribution function is the solution. We give a necessary and sufficient condition for singularity. We also show that they have a relationship with stationary measures.     

First and second order linear dynamic equations on time scales

We consider first and second order linear dynamic equations on a time scale. Such equations contain as special cases differential equations, difference equationsq— difference equations, and others. Important properties of the exponential function for a time scale are presented, and we use them to derive solutions of first and second order linear dyamic equations with constant coefficients. Wronskians are used to study equations with non—constant coefficients. We consider the reduction of order method as well as the method of variation of constants for the nonhomogeneous case. Finally, we use the exponential function to present solutions of the Euler—Cauchy dynamic equation on a time scale.     

Exponential mean-square stability of the improved split-step theta methods for non-autonomous stochastic differential equations

We consider the mean-square stability of the so-called improved split-step theta method for stochastic differential equations. First, we study the mean-square stability of the method for linear test equations with real parameters. When θ 3/2, the improved split-step theta methods can reproduce the mean-square stability of the linear test equations for any step sizes h 0. Then, under a coupled condition on the drift and diffusion coefficients, we consider exponential mean-square stability of the method for nonlinear non-autonomous stochastic differential equations. Finally, the obtained results are supported by numerical experiments.     

Synthesis of the Optimal Control of Linear Logical-Dynamical Systems Under Instant Multiple Switchings of the Automaton Part

We consider a dynamic system controlled by an automaton with memory. The continuous part of the system is described by linear differential equations and the logical (automaton) part is described by linear recurrence equations. The moments of the state change of the automaton part are not known in advance and they are determined in the process of optimization. Modes with instant multiple switchings of the automaton part are admitted. Based on sufficient optimality conditions, we develop a technique for the synthesis of a feedback control. The application of the technique is demonstrated in an example.     

Additive Schwarz algorithms for parabolic convection-diffusion equations

Summary In this paper, we consider the solution of linear systems of algebraic equations that arise from parabolic finite element problems. We introduce three additive Schwarz type domain decomposition methods for general, not necessarily selfadjoint, linear, second order, parabolic partial differential equations and also study the convergence rates of these algorithms. The resulting preconditioned linear system of equations is solved by the generalized minimal residual method. Numerical results are also reported.This work was supported in part by the National Science Foundation under Grant NSF-CCR-8903003 at the Courant Institute, New York University and in part by the National Science Foundation under contract number DCR-8521451 and ECS-8957475 at Yale University     

The conditioning of some projection methods for Fredholm and singular integral equations

We consider a general scheme for bounding the condition number of matrices arising from projection methods for solving linear operator equations. Applications are given for some Galerkin and collocation methods for Fredholm and Cauchy singular integral equations.     

Turning points of linear systems and the double asymptotics of the Painlevé transcendents

On the basis of the connection between the theories of linear and nonlinear special functions, we present a method which makes it possible to consider the well known formal limits from complicated Painlevé equations to simpler ones as the double asymptotics of specific solutions of these equations with respect to the parameter and the argument under some special relation between them. The hierarchies of the first and second Painlevé equations are interpreted as special functions that describe the isomonodromic collision of turning points for linear systems of ordinary differential equations. Bibliography: 28 titles. Translated fromZapiski Nauchnykh Seminarov POMI, Vol. 187, pp. 53–74, 1990. Translated by B. M. Bekker.     

Numerical solution and structural analysis of two-dimensional integral-algebraic equations

The υ-smoothing property of a one-dimensional Volterra integral operator and some projectors (Liang and Brumer, SIAM J. Numer. Anal. 51, 2238–2259 (2013)) are extended for two-dimensional integral-algebraic equations (TIAEs). Using these concepts, we decompose the given general TIAEs into mixed systems of two-dimensional Volterra integral equations (TVIEs) consisting of second- and first-kind TVIEs. Numerical technique based on the Chebyshev polynomial collocation methods is presented for the solution of the mixed TVIE system. Global convergence results are established and the performance of the numerical scheme is illustrated by means of some test problems.     

Constructing the gain matrix of a linear observer with given characteristics

We consider control systems with incomplete information about the phase state vector that are representable by linear difference equations. We solve the problem of algorithmic synthesis of a linear observer ensuring the required spectral properties for the state estimation error system.Kiev University. Translated from Vychislitel'naya i Prikladnaya Matematika, No. 67, pp. 140–146, 1989.