Regular Article: Computing Fourier Transforms and Convolutions on the 2-Sphere
Advances in Applied Mathematics
Journal of Complexity
Spherical Marcinkiewicz-Zygmund inequalities and positive quadrature
Mathematics of Computation
Local quadrature formulas on the sphere
Journal of Complexity
Optimal lower bounds for cubature error on the sphere S2
Journal of Complexity
Weighted quadrature formulas and approximation by zonal function networks on the sphere
Journal of Complexity
Cubature over the sphere S2 in Sobolev spaces of arbitrary order
Journal of Approximation Theory
Quadrature in Besov spaces on the Euclidean sphere
Journal of Complexity
Integral operators on the sphere generated by positive definite smooth kernels
Journal of Complexity
Optimal lower bounds for cubature error on the sphere S2
Journal of Complexity
Weighted quadrature formulas and approximation by zonal function networks on the sphere
Journal of Complexity
On the widths of Sobolev classes
Journal of Complexity
Complexity of numerical integration over spherical caps in a Sobolev space setting
Journal of Complexity
Numerical integration with polynomial exactness over a spherical cap
Advances in Computational Mathematics
Widths of Besov classes of generalized smoothness on the sphere
Journal of Complexity
Hi-index | 0.00 |
Let Sd-1 = {(x1,....,xd ∈ Rd : x12 + ...... + xd2 = 1} be the unit sphere of the d- dimensional Euclidean space Rd. For r Bpp (1 ≤ p ≤ ∞) the class of functions f on Sd-1 representable in the form f(x) = φ*Fr(x) = ∫Sd-1 φ(y)Fr(xy)dσ(y), ||φ||p ≤ 1, where dσ(y) denotes the usual Lebesgue measure on Sd-1, Fr(xy) = ∑ ∞ k=1 (k(k + 2λ))-r/2 γ(λ)(k+λ)/2πλ+1 Pkλ(xy), λ = d-2/2 and Pkλ(t) is the ultraspherical polynomial. For 1 ≤ p,q ≤ ∞, the Kolmogorov N-width of Bpr in Lq(Sd-1) is given by dN(Bpr, Lq) = XN supf∈Bprg∈XN||f - g||q, the left-most infimum being taken over all N-dimensional subspaces XN of Lq(Sd-1). The main result in this paper is that for r ≥ 2(d - 1)2, dN(Bpr, Lq) = = {N-r/d-1 if 2 ≤ p-r/d-1-1/p+1/2 if 1≤p≤q≤∞, where AN = NN means that there exists a positive constant C, independent of N, such that C-1 AN ≤ BN ≤ CAN. This extends the well-known Kashin theorem on the asymptotic order of the Kolmogorov widths of the Sobolev class of the periodic functions.