Nonlinear nonnested spline approximation (Q524407)

Linear and in particular non-linear spline approximation is a most useful tool in the approximation of for instance two-dimensional functions. Usually, piecewise polynomial splines with more and more refined knot-sequences are considered as elements of nested spaces spanned by splines. Generalising from this point of view, it is interesting to consider now non-nested spline spaces for the purpose of non-linear approximations of suitably smooth approximands. This approach is being followed in this article. Standard inequalities (that is, standard in the two-dimensional piecewise polynomial approach with nested subspaces) such as Jackson's and Bernstein's inequalities are established, for instance with the purpose of sharp inverse estimates for these spline approximations (direct and inverse estimates are proven). The approximations are considered in \(L^p\) and Besov spaces. The supports of the splines have the geometry of (convex) rings in two-dimensions. No polygonal boundaries are required for these rings.