关于苏联科学院数学研究所在函数逼近论方面的工作（上）

本文全面介绍了苏联科学院数学研究所的逼近论专家们多年来在函数逼近论的各个领域内所做的工作，这些领域包括：多项式的极值性质，一元周期函数与非周期函数逼近的正逆定理，线性逼近方法，多元函数逼近，函数类的宽度和求积公式等，在叙述这些成果的同时，作者还介绍了前人的有关结果．     

多元函数的最佳逼近递减的逼近阶

关于找一个必要与充分条件使得ωｉ１…ｋｉｍ（ｆ,１σｉ１,…,１σｉｍ）～Ｙσｉ１…σｉｍ（ｆ）,（σｉｊ→∞）成立的Ｔｉｍａｎ问题被解决。这条件是ωｋ１…ｋｉｍ（ｆ,δｉ１…δｉｍ）～ωｋｉ１＋１,…,ｋｉｍ＋１（ｆ,δｉ１…δｉｍ）,（δｉｊ→０）。     

单个函数平移和伸缩的叠加在L^p（R^d）中的逼近

本文主要讨论单个函数平移和伸缩的线性组合对Ｌ＾ｐ（Ｒ＾ｎ）中一紧集内函数的逼近，给出一个很强的逼近结果，这在神经网络应用研究中具有重要意义。     

绝对值函数的一致光滑逼近函数

绝对值函数是一个非光滑函数,研究了绝对值函数的光滑逼近函数.给出了绝对值函数的上方一致光滑逼近函数和下方一致光滑逼近函数,分别研究了其性质,并通过图像展示了逼近效果.     

在W2^1（R）空间中函数逼近的一种新方法

在再生核空间W2^1(R)中，利用再生核的性质实现了既不用计算导数也不需要计算积分，而只用函数值就可以将函数展开成级数的一种方法，并且这种级数的部分和{fn(x)}作为逼近f(x)的序列，它的误差rn(x)=f(x)-fn(x)在空间范数意义下单调下降．     

论可微函数的共单调逼近和共凸逼近

对有限区间上可微函数借助于代数多项式的共单调逼近和共凸逼近的逼近度估计建立了更为精确的Jackson型不等式,扩充和改进了近期的一些结果。     

关于一类函数的代数逼近度的注记

设A（z)是函数列{fn(z)}的极限函数，已知^[4],当|fn(z)-A(z)|在z=0的阶满足一定的条件时，A（z)是超越函数。本文给出了极限函数的代数逼近度函数。     

一类线性正算子对多元函数的点态逼近度

本文将所谓的ＳＢＫ算子推广为更为一般的多项式算子，研究了它对所谓Ｂ－有界变差的多元函数的点态逼近，改进并推广了文［６］和文［７］的结果。     

Approximation of anisotropic classes by wavelets

In this paper, we study the best approximation for anisotropic Sobolev and Besov classes in the Lq(Rd) metric by wavelets and obtain some asymptotic estimates of approximation order.     

Approximation of semigroups and cosine functions in spaces of periodic functions

In this article we furnish a representation of the solutions of some classes of first-order and second-order evolution problems as limit of iterates of classical sequences of approximating operators. The method is based on Trotter's theorem on the approximation of semigroups which is applied here also for the approximation of groups and cosine functions. We apply this method in spaces of continuous periodic functions and using some classical sequences of trigonometric polynomials.     

Orlicz空间中三角多项式倒数对周期可微函数的逼近定理

利用Orlicz空间内有关不等式技巧在Orlicz空间内研究了用三角多项式的倒数逼近周期可微函数的问题.得到了一个逼近定理及其推论.     

On Approximation by Ridge Functions

Ridge functions are defined as functions of the form , where , belongs to the given ``direction' set . In this paper we study the fundamentality of ridge functions for variable directions sets A and discuss the rate of approximation by ridge functions. Date received: June 7, 1994. Date revised: August 3, 1995.     

On Approximation by Reciprocals of Spherical Harmonics in L p Norm

Let S^1-1,q≥2,be the surface of the unit sphere in the Euclidean space R^1,f（x）∈L^p（S^q-1）,f（x）≥0,f absohutely unegual to 0,1≤p≤＋∞,Then,it is proved in the present paper that there is a spherical harmonics PN（x） of order≤N and a constant C〉0 such that where ω（f,δ）L^p=sup 0〈t≤δ‖St（f）-f‖L^p is a kind of moduli of continuity and ^‖f-1/PN‖L^p≤Cω（f,N^-1）L^p,St（f,μ）=1/｜S^q-2｜Sin^2λt ∫-μμ’=t f（μ＇）dμ＇ is a translation operator.     

关于周期神经网络逼近阶的研究

借助于有关Fourier级数的Riesz平均构造出了一类含有一个隐含层的周期神经网络与平移网络,与已有的讨论相比较,在获得相同的逼近阶的情况下,此类网络的隐层单元要求较少的神经元个数.     

单位根上重Hermite插值多项式的逼近阶

设ｆ（ｚ）在｜ｚ｜≤１解析，在｜ｚ｜≤１连续．本文得到了基于单位根的扩充Ｈｅｒｍｉｔｅ插值多项式在｜ｚ｜≤１上一致收敛于ｆ（ｚ）的逼近阶和在｜ｚ｜＝１上平均收敛于ｆ（ｚ）的逼近阶，且一般说来，阶还是精确的．进而说明，重数不同的插值多项式的逼近阶不同于重数相同的插值多项式的逼近阶．     

The Order of Best Approximation in Some Classes of Functions

Starting from the equivalence between the Ditzian–Totik modulus and , where , in this article large classes of functions are introduced for which the modulus can be easily calculated. As a consequence, very good estimates for the bestapproximation are obtained. The attempts to estimate or calculate themodulus can be a very intricateproblem.     

Approximation of Classes of Periodic Functions in Several Variables

We study the approximation of the classes and of periodic functions of several variables by multiple Fourier sums of fixed order constructed with regard to individual properties of functions from these classes. In a number of cases, such approximations allow us to achieve a better degree of approximation of the classes indicated above as compared to their approximation by staircase hyperbolic Fourier sums.     

Kinetic Theory of Quantum Electrodynamic Plasma in a Strong Electromagnetic Field: II. The Covariant Mean-Field Approximation

A covariant kinetic equation for the matrix Wigner function is derived in the mean-field approximation from a general kinetic equation for the fermionic subsystem of a quantum electrodynamic plasma. We show that in the semiclassical limit, the equations for the components of the Wigner function can be transformed into closed kinetic equations for the Lorentz-invariant distribution functions of particles and antiparticles.