Metric Approximation of Set-Valued Functions of Bounded Variation by Integral Operators

Several types of ℋ︁‐matrices were shown to provide a data‐sparse approximation of non‐local (integral) operators in FEM and BEM applications. The general construction is applied to the operators with asymptotically smooth kernel function provided that the Galerkin ansatz space has a hierarchical structure.The new class of ℋ︁‐matrices is based on the so‐called blended FE and polynomial approximations of the kernel function and leads to matrix blocks with a tensor‐product of block‐Toeplitz (block‐circulant) and rank‐k matrices. This requires the translation (rotation) invariance of the kernel combined with the corresponding tensor‐product grids. The approach allows for the fast evaluation of volume/boundary integral operators with possibly non‐smooth kernels defined on canonical domains/manifolds in the FEM/BEM applications. (Here and in the following, we call domains canonical if they are obtained by translation or rotation of one of their parts, e.g. parallelepiped, cylinder, sphere, etc.) In particular, we provide the error and complexity analysis for blended expansions to the Helmholtz kernel. Copyright © 2002 John Wiley & Sons, Ltd.

Local smoothness of functions and Bernstein-Durrmeyer operators

The purpose of this paper is to construct an integral rational Fourier operator based on the system of Chebyshev – Markov rational functions and to study its approximation properties on classes of Markov functions. In the introduction the main results of well-known works on approximations of Markov functions are present. Rational approximation of such functions is a well-known classical problem. It was studied by A. A. Gonchar, T. Ganelius, J.-E. Andersson, A. A. Pekarskii, G. Stahl and other authors. In the main part an integral operator of the Fourier – Chebyshev type with respect to the rational Chebyshev – Markov functions, which is a rational function of order no higher than n is introduced, and approximation of Markov functions is studied. If the measure satisfies the following conditions: suppμ = [1, a], a > 1, dμ(t) = ϕ(t)dt and ϕ(t) ἆ (t − 1)α on [1, a] the estimates of pointwise and uniform approximation and the asymptotic expression of the majorant of uniform approximation are established. In the case of a fixed number of geometrically distinct poles in the extended complex plane, values of optimal parameters that provide the highest rate of decreasing of this majorant are found, as well as asymptotically accurate estimates of the best uniform approximation by this method in the case of an even number of geometrically distinct poles of the approximating function. In the final part we present asymptotic estimates of approximation of some elementary functions, which can be presented by Markov functions.

- In this paper we consider the ill-posed problem of recovery of a function which has a finite number of jumps using noised data. The method which we suggest allows us to find a number of discontinuities. The position of the discontinuity points and the values of jumps are derived approximately. The estimates of accuracy of positions and jumps values are given. Approximation of initial function is constructed. The uniform estimate of proximity of the approximate function and the exact function is given outside the neighborhoods discontinuity points.

Dirac’s formalism combined with complex Fourier operational matrices to solve initial and boundary value problems

The convergence of iterated Boolean sums for different sequences of operators has been studied keeping fixed the number M of iterations or recently letting for a fixed operator. Here we consider some convergence properties of iterated Boolean-type sums of the classical Bernstein and Bernstein–Durrmeyer operators in the case where both the operators and their order of iterations are not fixed. In the case of Bernstein–Durrmeyer operators, we obtain a general solution of this problem. We can also state some alternative expressions of the semigroups and resolvent operators associated with the Voronovskaja’s formula for these sequences of operators.

We estimate pointwise convergence rates of approximation for functions of bounded variation and for functions which are exponentially bounded and locally of bounded variation. The approximation is through the operation of a sequence of integral operators with not necessarily positive kernel functions. The general result is then applied to deduce estimates for particular operators, such as Beta operators, Fourier–Legendre operators, Picard operators, and Gauss–Weierstrass operators.

This paper is devoted to some Korovkin approximation results in cones of Hausdorff continuous set-valued functions and in spaces of vector valued functions. Some classical results are exposed in order to give a more complete treatment of the subject. New contributions are concerned both with the general theory than in particular with the so-called convexity monotone operators, which are considered in cones of set-valued function and also in spaces of vector-valued functions.

In this paper, we prove that if a multilinear operator $\mathcal{T}$ and its multilinear commutator $\mathcal{T}_{\Sigma\vec{b}}$ and iterated commutator $\mathcal{T}_{\Pi\vec{b}}$ for $\vec{b}\in(\mathbb{R}^n)^m$ are bounded on product weighted Lebesgue space, then $\mathcal{T}$, $\mathcal{T}_{\Sigma\vec{b}}$ and $\mathcal{T}_{\Pi\vec{b}}$ are also bounded on product weighted Amalgam space. As its applications, we show that multilinear Littlewood-Paley functions and multilinear Marcinkiewicz integral functions with kernels of convolution type and non-convolution type, and their multilinear commutators and iterated commutators are bounded on product weighted Amalgam space. We also consider multilinear fractional type integral operators and their commutators' behaviors on weighted amalgam space. In order to deal with the endpoint case, we introduce the amalgam-Campanato space and show that fractional integral integral operator are bounded operators from product amalgam space to amalgam-Campanato space. What should point out is that even if in the linear case, our results for fractional integral operator are also new.

Under certain conditions, we obtain global pointwise estimates for Neumann series associated with an integral operator with general quasimetric kernel. The estimates involve the first and second iterations of the original kernel. As a consequence we deduce sharp bilateral bounds of Green's function for the fractional Schrödinger operator (− Δ)α/2 −q with general attractive potential q ≥ 0 on the entire Euclidean space \(\mathbb{R}^n\) for 0 < α < n, or a bounded nontangentially accessible domain \(\Omega \subseteq \mathbb{R}^n \) for 0 < α ≤ 2, under a certain “smallness condition” on q. Most of our results are new even in the classical case α = 2.KeywordsBounded DomainMeasure SpaceLipschitz DomainEuclidean BallDoubling MeasureThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

It is shown that the problem of the best uniform approximation in the Hausdorff metric of a continuous set-valued map with finite-dimensional compact convex images by constant set-valued maps whose images are balls in some norm can be reduced to a visual geometric problem. The latter consists in constructing a spherical layer of minimal thickness which contains the complement of a compact convex set to a larger compact convex set.

Nyström discretization of parabolic boundary integral equations

Many topological spaces are simply sets of points (atoms) endowed with a topology. Some spaces, however, have elements that are functions, matrices, or other non-atomic items. Another special type of space has elements that are themselves subsets of another space; these spaces are called hyperspaces. Hyperspaces are metric spaces, and the metric defined on them is called the Hausdorff metric. Hyperspaces wllose points are the closed subsets and hyperspaces whose points are the closed connected subsets (of metric spaces) have been extensively studied. Also, hyperspaces of closed and convex subsets of a bounded convex set in Euclidean space are of great interest in geometry. See Lay [3]. In recent years, geometers and topological dynamicists have explored spaces of closed and bounded subsets of the plane in connection with the study of fractals. One of the major results in the theory is that the hyperspace of closed subsets of a closed interval of real numbers is homeomorphic with the Hilbert cube, and with the space IX, the countable product of unit intervals. For more general information about the Hausdorff metric and spaces of subsets, see Devaney [2] and Sieradski [6]. The main result of this paper is a topological characterization of the space of closed subsets of a convergent sequence of points. The proof given here provides a homneornorphic embedding of the space in the plane, E2. The result was first proved by Pelczynski [5]. He studied arbitrary compact zero-dimensional metric spaces, so his proofs are much more widely applicable, but they are also somewhat technical. Our proof depends only on some well-known results in the theory of metric spaces, and is therefore accessible to advanced undergraduate mathematics majors. We also prove that for no convergent sequence of real numbers is there an isometric embedding of the hyperspace in Euclidean space, E,1, for any n. For other results and discussion, see Nadler [4]. Let (X, d) denote a compact metric space. The hyperspace (2X, D) of X is the metric space whose points are the closed nonempty subsets of X and whose metric is the Hausdorff metric D given by

Diffuse Optical Tomography (DOT) involves a nonlinear optimization problem to find the tissue optical properties by measuring near-infrared light noninvasively. Many researchers used linearization methods to obtain the optical image in real time. However, the linearization procedure may neglect small but sometimes important regions such as small tumors at an early stage. Therefore, nonlinear optimization methods such as gradient- or Newton- type methods are exploited, resulting in better resolution image than that of linearization methods. But the disadvantage of nonlinear methods is that they need much computation time. To solve this trade-off dilemma between image resolution and computing time, we suggest second order inverse Born expansion algorithm in this paper. It is known that a small perturbation of photon density is represented by Born expansion with respect to the perturbation of optical coefficients, which is an infinite series of integral operators having Robin function kernel. Whereas, inverse Born expansion is an implicit representation of a small perturbation of optical coefficients by an infinite series of the integral operators with respect to the photon density and its perturbation, which is appropriate series expansion for inverse DOT problem. Solving the inverse Born expansion itself and the first order approximation correspond to nonlinear and linear method, respectively. We formulated a second order approximation of the inverse Born expansion explicitly to make numerical implementation possible and showed the convergence order of the proposed method is higher than the linear method.

The aim of this paper is to introduce a concept of quasiconvexity for set-valued maps in a general framework, by only considering an abstract convexity structure in the domain and an arbitrary binary relation in the codomain. It is shown that this concept can be characterized in terms of usual quasiconvexity of certain real-valued functions. In particular, we focus on cone-quasiconvex set-valued maps with values in a partially ordered vector space.