Convergence of Newton's method for convex best interpolation (Q5927645)

The authors consider the following convex best interpolation problem: find a convex function which interpolates given points and has a minimal \(L^2\) norm of the second derivative. An extension of the Newton method, the so-called semismooth generalized Newton method of \textit{L. Qi} and \textit{J. Sun} [Math. Programming 58A, No. 3, 353-367 (1993; Zbl 0780.90090)] is analysed and its numerical efficiency is confirmed by numerical experiments.