scientific article; zbMATH DE number 1042858

From MaRDI portal
Publication:4346407





Earlier the authors proposed locally optimal splines as a special solution of the problem of approximation with a minimal number of knots. A locally optimal spline is a spline with a minimal number of knots with a preassigned accuracy only for convex functions. On the basis of special alternance properties of locally optimal splines two algorithms for their construction were proposed. The first algorithm is based on the use of locally supporting rays to the boundaries of an admissibility domain with an isolation of a ray of maximal length. The second algorithm is based on the alternance properties of a chain of a locally optimal spline. The restriction on the convexity of the function being approximated was removed by the introduction of left-optimal (right-optimal) spline, which are used to construct a globally optimal spline. An algorithm for constructing an optimal spline was developed with the use of the properties of left-optimal (right-optimal) splines. In this paper we generalize the studies of locally optimal splines and Chebyshev optimal splines. The final classification of types of solution to the Chebyshev inverse problem on the class of piecewise linear splines and the corresponding terminology are given elsewhere. For the entire collection see [Zbl 0863.00015].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4346407)