Approximate Recovery of Multivariate Periodic Functions using Wavelet Decompositions

Dinh Dung and Mai Xuan Thao

Abstract

     We study the optimal recovery of multivariate periodic functions of the Besov class of common smoothness SBwp,q  from their values at  n points in terms of the quantity Rn(SBwp,q ,Lq), which is a characterization of optimality of methods of recovery. The smoothness of SBwp,q  is defined via modulus of smoothness dominated by a function w  of modulus of smoothness type. With some restrictions on w and p, q we give the asymptotic order of this quantity when n ---> ¥ . An asymptotically optimal method of recovery is constructed by using the wavelet family formed from the integer translates of the dyadic scales of multivariate de la Vallée Poussin kernels.