Approximation properties of periodic multivariate quasi-interpolation operators (Q822732)

Quasi-interpolation is one of the most useful approaches to the approximation of functions and data. Its manifold advantages are (i) its versatility due to the relatively simple formulation in the absense of for instance the standard interpolation conditions, (ii) its availability in many dimensions and its (iii) suitable properties such as smoothness and localness, not to forget (iv) the available efficient computational schemes for evaluations. Much of the importance of quasi-interpolation operators comes from the relatively free choice of ''quasi-Lagrange'' functions which form the bases for making the approximation. It may be a classical Lagrange function, it can be made to decay at various rates, the quasi-interpolation may reproduce (trigonometric) poylnomials etc. In this article, the authors consider \(d\)-variate quasi-interpolation to periodic functions. They are taken from linear spaces of trigonometric polynomials, and the information from the approximands may be provided by distributions applied to the functions. This includes of course function evaluation (and may include Lagrange interpolation if the basis functions [quasi-Lagrange functions] are formed as true Lagrange functions). The results provided give upper and lower error bounds on the quasi-interpolation error in \(L_p\)-norms.