Locally linearly independent bases for bivariate polynomial spline spaces (Q5934296)

For a regular triangulation \(\Delta\), let \(S^r_d(\Delta): =\{s \in C^r(\Omega): s\mid T\in\Pi_d\) for all \(T\in\Delta\}\), where \(\Pi_d\) denotes the set of bivariate polynomials of degree \(\leq d\) and \(\Omega\) is the union of all triangles \(T\in\Delta\). Suppose that \(\{B_i\}^n_{i=1}\) is a basis for \(S^r_d (\Delta)\). A basis for \(S^r_d(\Delta)\) is called locally linearly independent, if for every \(T\in\Delta\), the basis splines \(B_i\) \((i\in \Sigma_T)\) are linearly independent on \(T\), where \(\Sigma_T: =\{i:T \subseteq \text{supp} B_i \}\). Locally linearly independent bases play an important role in the theory of interpolation by multivariate splines. The authors construct locally linearly independent bases for \(S^r_d (\Delta)\) for all \(d\geq 3r+2\), and for the scale of superspline spaces. The construction is based on Bernstein-Bézier technique.