On lower bounds for the approximation of functions by local splines with nonfixed nodes

For functions integrable to the power , we obtain asymptotically exact lower bounds for the approximation by local splines of degree r and defect k< r/2 in the metric of L _p      

On approximation of functions from below by splines of the best approximation with free nodes

Let M be the set of functions integrable to the power β=(r+1+1/p)^-1. We obtain asymptotically exact lower bounds for the approximation of individual functions from the set M by splines of the best approximation of degree rand defect k in the metric of L _p.     

An approximation of analytical functions by local splines

Local splines are presented for the approximation of functions of one and many variables, which are analytic in the domains , where U_i(z_i) is a unit disk in the complex plane C_i,i=1,2,…,l, l=1,2, …. Results are given for functions whose r-order derivatives belong to the Hardy's class H_p,1≤p≤∞. It is shown that the approximation converge to the function at the rate for functions of one variable and An^{−(r−1/p)/(l−1)} for functions of l variables, where n is the number of points of local splines and A and C are positive constants. This work was supported by Russian Foundation of Fundumental Inverstigations     

Exact bounds for the uniform approximation of continuous periodic functions by r-th order splines

We solve the problem of determining exact bounds for the uniform approximation of continuous periodic functions by r-th order interpolation splines in a space C and on a class H specified by the convex modulus of continuity(t).Translated from Matematicheskie Zametki, Vol. 13, No. 2, pp. 217–228, February, 1972.In conclusion the author wishes to express his deep gratitude to N. P. Korneichuka for constant attention and observations which were useful to him in preparing the paper.     

Local approximation by splines with displacement of nodes

We consider the problem of approximating a function defined on a uniform mesh by the method of local polynomial spline-approximation where the mesh of the nodes of the spline is chosen displaced relative to the mesh of the initial data. Conditions are established for the local form preservation by the spline of the initial data.We study the approximative properties of the method for the case of the simplest local approximation formula and find the optimal values of the displacement parameters.     

Error bounds for the approximation of Green's kernels by splines

Summary In this paper we give error bounds for the approximation by tensor-product splines of surfaces which are defined on a square and which are smooth except along the diagonal.Supported in part by AFOSR Grant 77-3150     

Exact bounds for simultaneous approximation of functions of two variables and their derivatives by bilinear splines

We find the exact value of the expression $$\varepsilon ^{(l,q)} {\mathbf{ }}(W^{(r,s)} ){\mathbf{ }}H^{w_1 ,w_2 } (G)) = \sup \{ ||f^{(l,q)} ( \cdot {\mathbf{ }}, \cdot ) - S_{1,1}^{(l,q)} (f;{\mathbf{ }} \cdot {\mathbf{ }}, \cdot )||_{C(G)} :f \in W^{(r,{\mathbf{ }}s)} H^{w_1 ,w_2 } (G)\} ,$$ , where? ^(l,q) (x,y)=? ^1+q ?/?x ^l ?y ^q (l, q=0, 1, 1≤l+q≤2) andS _1,1(f; x, y) is a bilinear spline interpolatingf(x, y) in the nodes of the grid Δ _mn =Δ _m ^x ×Δ _n ^y with Δ _m ^x :x _i =i/m (i=0, ..., m) and Δ _n ^y :y _j =j/n (j=0, ..., n). Here $(W^{(r,s)} ){\mathbf{ }}H^{w_1 ,w_2 } (G)$ is the class of functionsf(x, y) with continuous derivativesf ^(r,s)(x, y) (r, s=0, 1, 1≤r+s≤2) on the squareG=[0, 1]×[0, 1] and with the modulus of continuity satisfying the inequalityω(f ^(r,s);t, τ)≤ω ₁ (t)+ω ₂ (τ), whereω ₁ (τ) andω ₂ (τ) are the given moduli of continuity.     

On best error bounds for approximation by piecewise polynomial functions

Summary An analog of the well-known Jackson-Bernstein-Zygmund theory on best approximation by trigonometric polynomials is developed for approximation methods which use piecewise polynomial functions. Interpolation and best approximation by polynomial splines, Hermite and finite element functions are examples of such methods. A direct theorem is proven for methods which are stable, quasi-linear and optimally accurate for sufficiently smooth functions. These assumptions are known to be satisfied in many cases of practical interest. Under a certain additional assumption, on the family of meshes, an inverse theorem is proven which shows that the direct theorem is sharp.The work presented in this paper was supported by the ERDA Mathematics and Computing Laboratory, Courant Institute of Mathematical Sciences, New York University, under Contract E(11-1)-3077 with the Energy Research and Development Administration.     

On error bounds for lower semicontinuous functions

We give some sufficient conditions for proper lower semicontinuous functions on metric spaces to have error bounds (with exponents). For a proper convex function f on a normed space X the existence of a local error bound implies that of a global error bound. If in addition X is a Banach space, then error bounds can be characterized by the subdifferential of f. In a reflexive Banach space X, we further obtain several sufficient and necessary conditions for the existence of error bounds in terms of the lower Dini derivative of f. Received: April 27, 2001 / Accepted: November 6, 2001?Published online April 12, 2002     

On lower bounds in radial basis approximation

We investigate the approximation by manifolds _n() generated by linear combinations of n radial basis functions on R^d of the form (|–a|), where is the thin-plate spline type function. We obtain exact asymptotic estimates for the approximation of Sobolev classes W^r(B^d) in the space L(B^d) on the unit ball B^d. AMS subject classification 41A25, 41A63, 65D07, 41A15     

Upper bounds for the best one-sided approximation by splines of the classes WrL1

In the present note we will investigate the problem of the one-sided approximation of functions by n-dimensional subspaces. In particular, we will find the exact value of the best one-sided approximation of the class W^rL₁ (r=1, 2, ...) of all periodic functions f(x) of period 2 for which f^(r–1)(x) (f⁽⁰⁾(x)=f(x)) is absolutely continuous and f^(r)L₁1 by periodic spline functions S_2n ( = 0, 1, ..., n=1, 2, ...) of period 2, order ,and deficiency 1.Translated from Matematicheskie Zametki, Vol. 19, No. 1, pp. 11–17, January, 1976.     

On the best approximation of periodic functions of two variables by polynomial splines

We consider the problem of the best approximation of periodic functions of two variables by a subspace of splines of minimal defect with respect to a uniform partition.     

On lower bounds for integration of multivariate permutation-invariant functions

In this note we study multivariate integration for permutation-invariant functions from a certain Banach space _Ed,α$E_{d,\alpha }$ of Korobov type in the worst case setting. We present a lower error bound which particularly implies that in dimension d$d$ every cubature rule which reduces the initial error necessarily uses at least d+1$d+1$ function values. Since this holds independently of the number of permutation-invariant coordinates, this shows that the integration problem can never be strongly polynomially tractable in this setting. Our assertions generalize results due to Sloan and Wo?niakowski (1997) [3]. Moreover, for large smoothness parameters α$\alpha$ our bound cannot be improved. Finally, we extend our results to the case of permutation-invariant functions from Korobov-type spaces equipped with product weights.     

On the order of relative approximation of classes of differentiable periodic functions by splines

In the case where n → ∞, we obtain order equalities for the best L _q -approximations of the classes W _p ^r , 1 ≤ q ≤ p ≤ 2, of differentiable periodical functions by splines from these classes.     

On extremal subspaces and approximation of periodic functions by splines of minimal defect

. . , 320 625 . 72      

Uniform approximation of min/max functions by smooth splines

In some optimization problems, min(f₁,…,f_n) usually appears as the objective or in the constraint. These optimization problems are typically non-smooth, and so are beyond the domain of smooth optimization algorithms. In this paper, we construct smooth splines to approximate min(f₁,…,f_n) uniformly so that such optimization problems can be dealt with as smooth ones.