Generalized polynomials of minimal norm

Approximation by polynomials (41A10) Approximation by other special function classes (41A30) Saturation in approximation theory (41A40) Best approximation, Chebyshev systems (41A50) Approximate quadratures (41A55) Abstract approximation theory (approximation in normed linear spaces and other abstract spaces) (41A65)

Review text

The authors consider the problem of finding optimal generalized polynomials of minimal \(L_ p\) norm, \(1\leq p<\infty\). Let \(\{u_ i\}^{N+1}_{i=1}\subset C^ N[a,b]\), [0,1]\(\subset (a,b)\), be a given set of functions with the property that \(\{u_ i\}^ K_{i=1}\) forms an ``extended Chebyshev system for \(K=N-1,N,N+1\). For \({\bar \tau}=(\tau_ 1,...,\tau_ N)\in R^ N\) the symbol u(\({\bar \tau}\);t) is devoted to denote the unique polynomial of the type \(u=u_{N+1}+\sum^{N}_{i=1}\alpha_ iu_ i\) with real coefficients \(\{\alpha_ i\}\) such that \(u(\tau_ k)=0\), \(k=1,...,N\). Let us consider the multiplicities \(\{\nu_ k\}^ n_{k=1}\), \(\{\mu_ k\}^ m_{k=1}\) such that \(\sum^{n}_{k=1}\nu_ k+\sum^{m}_{k=1}\mu_ k=N\) and the points \({\bar \xi}=(\xi_ 1,...,\xi_ m)\) where \(a<\xi_ 1\leq...\leq \xi_ m<b\). Set \(\Delta_ n=\{\bar x=(x_ 1,...,x_ n)\); \(0\leq x_ 1\leq x_ 2\leq...\leq x_ n\leq 1\}\) and for each \(\bar x=(x_ 1,...,x_ n)\in \Delta_ n\), let \({\bar \tau}={\bar \tau}(\bar x,{\bar \xi})\in R^ N\) be the form \[ {\bar \tau}(\bar x,{\bar \xi})=(x_ 1,...,x_ 1,x_ n,...,x_ n,\xi_ 1,...,\xi_ 1,\xi_ m,...,\xi_ m). \] If \(I(\bar x,{\bar \xi})=\int^{1}_{0}| u({\bar \tau}(\bar x,{\bar \xi});t)|^ p dt,\) \(1\leq p<\infty\), the aim of this important paper is to investigate the existence and the uniqueness of the \(\bar x^*\in \Delta_ n\) with the property \((1)\quad I(\bar x^*,{\bar \xi})=\min_{\bar x\in \Delta_ n}I(\bar x,{\bar \xi}).\) The main results are the following: ``Let \(1\leq p<\infty\). Then there is an \(\bar x^*\in \Delta_ n\) which satisfies (1). Further any such minimizing \(\bar x^*=(x^*_ 1,...,x^*_ n)\) is in interior of \(\Delta_ n\), i.e., \(0<x^*_ 1<...<x^*_ n<1\). For each n,m\(\geq 0\) so that \(n+m\leq N\), there is a unique solution \(\bar x^*\in \Delta_ n\) to the problem cited in (1). As an application of the above results, the authors obtain Gaussian quadrature formulas for extended Chebyshev systems.

Recommendations

Cites work

Cited in

(17)

This page was built for publication: Generalized polynomials of minimal norm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1822709)