scientific article; zbMATH DE number 15669

Interpolation in approximation theory (41A05) Approximation by polynomials (41A10) Orthogonal functions and polynomials, general theory of nontrigonometric harmonic analysis (42C05)

Review text

[For the entire collection see Zbl 0706.00007.] For any \(\{y_ k\}^ N_ 1\), \(-1\leq y_ 1<\cdots<y_ N<1\), let \(\omega(x)=\prod^ N_ 1| x-y_ k|^{\alpha_ k}\), \(\alpha_ k>-1\) be a generalized Jacobi weight (g.J.). If \(-1\leq x_ 1<\cdots<x_ n\leq 1\) are the zeros of the \(n\)-the orthogonal polynomial associated with \(\omega(x)\), let \(L_ n(f;x)\) denote the Lagrange interpolant to \(f\) on the nodes \(\{x_ i\}^ n_ 1\). The authors prove the following result which is related to the classical results of Erdős and Feldheim and Erdős and Turán on mean convergence. Theorem 1. If \(f^{(r)}\in C[-1,1]\), then for \(n\geq r+1\) \[ \begin{split} \int\limits^ 1_{-1}\left| f^{(r)}(x)-L^{(r)}_ n(f;x)\right|^ p(1- x^ 2)^{rp/2}\omega(x)dx \\ \leq Cn^{rp}\left(\int\limits^ 1_{- 1}\left| f(x)-L_ n(f;x)\right|^ p\omega(x)dx+\left(E_{n- r+1}(f^{(r)})\right)^ p\right),\end{split} \] (\(r=0,1,\dots;\) \(0<p\leq\infty\)) \((E_ n(f)\) denotes the error of least approximation by polynomials of degree \(\leq n\)). When \(r=2\), and \(\omega(x)=(1-x^ 2)^{-1/2}\), they prove a simpler inequality: \[ \int\limits^ 1_{- 1}\left\{f^{(r)}(x)-L^{(r)}_ n(f,x)\right\}^ 2\sqrt{1-x^ 2}dx \leq C_ r\left(\omega\left(f^{(r)},{1\over n}\right)\right)^ 2, \] where \(\omega\) denotes the modulus of continuity.

Recommendations

Cited in

(23)

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 Q3972313)