球面卷积算子的饱和阶

本文给出了确定球面卷积算子饱和阶的一般方法并举例说明了它的应用．     

学习理论中核函数逼近的Jackson型不等式（英文）

从微分算子角度理解核函数空间,借助经典Fourier变换研究核函数逼近问题.应用Fourier乘子算子和算子半群定义了一种光滑模,证明其与一种基于微分算子的K-泛函的等价性,由此给出了刻画核函数逼近收敛性的Jackson不等式.进一步证明,如果微分算子为Riesz势算子或Bessel势算子,逼近的收敛性可以转化为卷积算子逼近.特别地,给出了再生核Hilbert空间逼近的一种上界估计.     

球面神经网络的构造与逼近

研究球面神经网络的构造与逼近问题.利用球面广义的de la Vallee Poussin平均、球面求积公式及改进的单变量Cardaliaguet-Euvrard神经网络算子,构造具logistic激活函数的单隐层前向网络,并给出了Jackson型误差估计.     

由斜坡函数激发的神经网络算子逼近

首先,引入一种由斜坡函数激发的神经网络算子,建立了其对连续函数逼近的正、逆定理,给出了其本质逼近阶.其次,引入这种神经网络算子的线性组合以提高逼近阶,并且研究了这种组合的同时逼近问题.最后,利用Steklov函数构造了一种新的神经网络算子,建立了其在L~p[a,b]空间逼近的正、逆定理.     

球面上径向基函数最小范数插值逼近的误差估计

研究了球面径向基插值对球面函数的逼近问题,给出了一致逼近的上界估计式．文中结果说明,球面径向基插值的逼近阶会随函数光滑性的提高而增加．     

一类乘子算子在全测度集上的逼近

本文给出了研究乘子算子在全测度集上逼近的一种框架.在 Riesz 极大算子有界的条件下,确定了一类乘子算子在 Riesz 位势空间上几乎处处逼近的阶.并用于讨论广义 Bochner-Riesz 平均和 Abel-Cartwright 平均的点态逼近.     

球面广义移动最小二乘的diffuse泛函逼近

考虑单位球面上广义移动最小二乘的逼近问题.用一般意义的泛函代替标准移动最小二乘中的点值泛函,给出单位球面上广义移动最小二乘的定义,并在此基础上通过diffuse导数的概念定义一种diffuse泛函的移动最小二乘逼近,根据球面网格范数给出逼近阶估计.该diffuse泛函移动最小二乘逼近亦可被理解为一种"同时"逼近.     

一类多元Gauss-Weierstrass算子线性组合的逼近

本文主要讨论一类多元Gauss-Weierstrass算子的线性组合的逼近性质,建立了一致逼近下的正、逆定理,并给出了逼近阶的特征刻画.     

构造球面逼近算子的一种方法

借助于经典球面分析的Bochner-Riesz平均,Cesàro平均及有关球调和多项式的Gauss积分公式构造出了两类球面平移算子,并且以K-泛函为工具给出了逼近的上界估计.     

再论局部域上一类逼近恒同算子

在新近的文献[1]中我们给出局部域上一类逼近恒同算子f*K_ω(x),现继续这类算子的研究.主要讨论他们的逼近阶,逆逼近定理与相应的极大算子的型等.     

Convolution solutions of an abstract stochastic Cauchy problem

We study convolution solutions of an abstract stochastic Cauchy problem with the generator of a convolution operator semigroup. In the case of additive noise, we prove the existence and uniqueness of a weak convolution solution; this solution is described by a formula generalizing the classical Cauchy formula in which the solution operators of the homogeneous problem are replaced by the convolution solution operators of the homogeneous problem. For the problem with multiplicative noise, we find a condition under which the weak convolution solution coincides with the soft solution and indicate a sufficient condition for the existence and uniqueness of a weak convolution solution; the latter can be obtained by the successive approximation method.     

Degree of Lp-approximation by certain positive convolution operators

The degree of L_p-approximation for a class of positive convolution operators is investigated. Recent results of De Vore, Bojanic, and Shisha for the uniform approximation by these operators and the K-functional of Peetre are employed to obtain the degree of approximation in terms of the integral modulus of smoothness.     

On the cardinal spline interpolation corresponding to infinite order differential operators

This paper discusses some problems on the cardinal spline interpolation corresponding to infinite order differential operators. The remainder formulas and a dual theorem are established for some convolution classes, where the kernels arePF densities. Moreover, the exact error of approximation of a convolution class with interpolation cardinal splines is determined. The exact values of averagen-Kolmogorov widths are obtained for the convolution class. Supported in part by NSFC.     

Convergence and Rate of Approximation in BVΦ for a Class of Integral Operators

We obtain estimates and convergence results with respect to ?-variation in spaces BV_Φ for a class of linear integral operators whose kernels satisfy a general homogeneity condition. Rates of approximation are also obtained. As applications, we apply our general theory to the case of Mellin convolution operators, to that one of moment operators and finally to a class of operators of fractional order.     

The kernels of vallee poussin type on P-adic fields

In this paper, we present a class of approximation identity operators of convolution type on the p-acid field K and on the ring of integersD⊂K, and study their approximation properties. All the kernels of these operators are of approximation identity, and are called de la Vallée Poussin type (VP type kernels). The structure of the VP type kernels onD is essentially an analogue of the classical one, however, the approximation properties are much better. Additionally, we discuss the maximal VP type operators and their properties. Supported by National Science Foundation of China     

Classes of convolution type operators on unions of bounded intervals

An invertibility theory for classes of convolution type operators on a union of bounded intervals whose kernels have Fourier transforms which are related to solutions of corona problems is established and the corresponding formulas for the inverse operators are given. A generalization of the portuguese transformation for matrix functions is obtained and is used to establish the invertibility theory for one of the above mentioned classes of operators. The same transformation allows, also, to establish the equivalence between convolution type operators on an union of disjoint intervals and convolution type operators on a bounded interval.     

Convolution operators on spaces of entire functions

We show that nontrivial convolution operators on certain spaces of entire functions on E are frequently hypercyclic when E is a normed space and when E is the strong dual of a Fréchet nuclear space. We also obtain results of existence and approximation for convolution equations on certain spaces of entire functions on arbitrary locally convex spaces.     

The best approximation to a class of functions of several variables by another class and related extremum problems

We study the relationship between several extremum problems for unbounded linear operators of convolution type in the spaces $L_\gamma = L_\gamma (\mathbb{R}^m )$ , m ≥ 1, 1 ≤ γ ≤ ∞. For the problem of calculating the modulus of continuity of the convolution operatorA on the function classQ defined by a similar operator and for the Stechkin problem on the best approximation of the operatorA on the classQ by bounded linear operators, we construct dual problems in dual spaces, which are the problems on, respectively, the best and the worst approximation to a class of functions by another class.     

Bernstein型算子加Jacobi权逼近

对于Bernstein型算子,证明它在通常的加权范数下是无界的,通过引进新的加权范数,研究其加Jacobi权的逼近性质,得到加权逼近的正逆定理,从而导出加权逼近特征的等价刻画．