A parsimonious approach to $C^2$ cubic splines on arbitrary triangulations: Reduced macro-elements on the cubic Wang-Shi split
Tom Lyche, Carla Manni, Hendrik Speleers
TL;DR
This paper addresses constructing $C^2$ cubic splines on the $WS_3$ refinement of an arbitrary triangulation by concentrating Hermite degrees of freedom at vertices and edges, achieving a parsimonious representation via macro-elements. It develops reduced subspaces ${\mathbb S}_3^{2,m}({\mathcal T}_{WS_3})$ containing ${\mathbb P}_3$ and represented by a local simplex-spline basis, enabling per-triangle construction and straightforward global coupling. The approach yields subspaces with dimension up to six times the number of vertices and provides explicit conditions and procedures to preserve cubic polynomial reproduction while removing interior degrees of freedom. The resulting framework offers an efficient and geometrically robust method for $C^2$ cubic splines on complex triangulations, with practical implementation aided by the simplex-spline basis and the necessary basis-conversion matrices.
Abstract
We present a general method to obtain interesting subspaces of the $C^2$ cubic spline space defined on the cubic Wang-Shi refinement of a given arbitrary triangulation $\mathcal{T}$. These subspaces are characterized by specific Hermite degrees of freedom associated with only the vertices and edges of $\mathcal{T}$, or even only the vertices of $\mathcal{T}$. Each subspace still contains cubic polynomials while saving a consistent number of degrees of freedom compared with the full space. The dimension of the considered subspaces can be as small as six times the number of vertices of $\mathcal{T}$. The method fits in the setting of macro-elements: any function of such a subspace can be constructed on each triangle of $\mathcal{T}$ separately by specifying the necessary Hermite degrees of freedom. The explicit local representation in terms of a local simplex spline basis is also provided. This simplex spline basis intrinsically takes care of the complex geometry of the Wang-Shi split, making it transparent to the user.
