Cubic spline interpolation of functions with high gradients in boundary layers (Q2357115)

New spline approximation and interpolation methods are proposed to approximate functions whose gradients are steep in certain regions of a bounded set. Splines and cubic splines as a special case are often used in one or more dimensions with equally spaced grids, but these -- unsurprisingly -- are inefficient for the aforementioned situations. Non-equally spaced meshes and sharp error estimates are provided in this article to overcome these difficulties.