Cubic spline fitting using data dependent triangulations (Q2277754)

A kind of data dependent triangulation technique is effectively developed by the authors to investigate the data fitting problem: ``Given a set of points \({\mathcal P}=\{(x_ i,y_ i)\}^ n_{i=1}\) in \({\mathbb{R}}\) and a corresponding set of measurements \(\{z_ i=f(x_ i,y_ i)\}^ n_{i=1}\) on a function f, construct a surface s approximating f.'' They make use of piecewise cubic \(C^ 1\)-surfaces, and base the selection of the triangulation upon the idea of minimizing the energy of the resulting surface. Some explicit constructions and a pair of algorithms are given.