Ambient approximation on hypersurfaces (Q1755513)

In this article, a new general approach to solving approximation problems on hypersurfaces is presented. The approach draws on applying standard approximation techniques in the embedding space. Given certain mild conditions, the authors show that it is possible to transfer approximation orders from the embedding space onto the hypersurface. When applied to tensor-product B-splines of order \(n\in \mathbb{N}\) and knot spacing \(h > 0\), it is established that the resulting approximant is in \(C^{n-2}\) and has an optimal error decay of \(h^n\). Moreover, this new method is easy to implement and optimally efficient. Examples using smooth surfaces of arbitrary genus are presented.