Raising the approximation order of multivariate quasi-interpolants (Q466812)

In the context of approximations by the important ansatz of quasi-interpolation defined on domains in \(s\) dimensional ambient spaces, finite sets (bases) of functions are used to construct approximations to smooth functions. Typical applications of the setting described are the approximations by spline spaces, i.e., local approximations by piecewise polynomials defined on -- say -- two-dimensional triangulations. The spaces are supposed to contain certain polynomials at a minimum, and the approximants are then created using the approach of quasi-interpolation. This method is a highly useful one and often quasi-interpolation is indeed the most suitable type of approximation used in technical and scientific applications. The main results of the article demonstrate how the classical approximation orders of quasi-interpolants may be improved for this setting.