Best L_ p approximation with multiple constraints for 1 p<

Spline approximation (41A15) Approximation with constraints (41A29) Best approximation, Chebyshev systems (41A50) Spaces of measurable functions ((L^p)-spaces, Orlicz spaces, Köthe function spaces, Lorentz spaces, rearrangement invariant spaces, ideal spaces, etc.) (46E30)

Review text

A real-valued function g is said to be n-convex in (0,1) if for any \(n+1\) distinct points \(x_ 0,...,x_ n\) in (0,1), the nth order divided difference is nonnegative, i.e. \([x_ 0,x_ 1,...,x_ n]g\geq 0.\) Given \(0\leq m\leq n\), g is said to be (m,n)-convex if \((-1)^ ig\) is \((m+i)\)- convex for \(i=0,...,n-m\). Let \(K_{m,n}\) denote the set of (m,n)-convex and \(K^ p_{m,n}=K_{m,n}\cap L_ p[0,1].\) Given a partition \(\Delta\) of [0,1], with \(\Delta:0=x_ 0<x_ 1<...<x_{n+1}=1\), let \(S^ k_ n(\Delta)\) denote the space of polynomial splines of degree n- 1 with k simple knots at \(x_ 1,...,x_ k\) and \(S^ k_{m,n}(\Delta)=S^ k_ n(\Delta)\cap K_{m,n}.\) Given \(f\in L_ p[0,1]\), \(S^*\in K^ p_{m,n}\) (resp., \(S^ k_{m,n}(\Delta))\) is called a best (m,n)-convex (resp., (m,n)-convex spline) \(L_ p\) approximation to f if \(\| f-S^*\| p=\inf \{\| f-S\|_ p,\) \(s\in K^ p_{m,n}\) (resp., \(S^ k_{m,n}(\Delta))\}\). The problem considered in this paper is best \(L_ p\) \((1\leq p<\infty)\) approximations from \(S^ k_{m,n}(\Delta)\) and \(K^ p_{m,n}\). Characterizations of best \(L_ p\) approximations from \(S^ k_{m,n}(\Delta)\) and \(K^ p_{m,n}\) are established and the relationship between them is investigated.

Recommendations

Cites work

Cited in

(7)

This page was built for publication: Best \(L_ p\) approximation with multiple constraints for \(1\leq p<\infty\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q757769)