Closed balls for interpolating quasi-polynomials (Q430538)

This interesting paper studies interpolation in spatial points using so called `quasi-polynomials' of the form \[ {\mathbf f}(u)={\mathbf b}_1u^{\alpha_1}+{\mathbf b}_2u^{\alpha_2}+\cdots+{\mathbf b}_nu^{\alpha_n},\;u\in [0,\infty), \eqno{(\ast)} \] where \({\mathbf b}_1,{\mathbf b}_2,\dots{\mathbf b}_n\in{\mathbb C},\alpha_1,\alpha_2,\dots,\alpha_n\in{\mathbb R}\) with \(0=\alpha_1<\alpha_2<\cdots<\alpha_n\). The interpolation problem consists of finding a function of the form \((\ast)\) for any given set of data pairs \((x_1,{\mathbf y}_1), (x_2,{\mathbf y}_2),\dots,(x_n,{\mathbf y}_n)\), where \(0\leq x_1<x_2<\cdots<x_n\), and \({\mathbf y}_1,{\mathbf y}_2,\dots, {\mathbf y}_n\in{\mathbb C}\), such that \[ {\mathbf f}(x_i)={\mathbf y}_i,\;i=1,2,\dots,n. \] In a paper by \textit{J. Wen} and \textit{W.-L. Wang} [Comput. Math. Appl. 56, No. 4, 1045--1058 (2008; Zbl 1155.26310)], this type of problem has been studied and several properties were derived, once existence/uniqueness of the solution is shown. The paper under review specifically studies the problem of, given a spatial curve \({\mathbf g}\) in \({\mathbb R}^m\) by information in a set of data points \((x_1,{\mathbf y}_1), (x_2,{\mathbf y}_2),\dots,(x_n,{\mathbf y}_n)\), where \(0\leq x_1<x_2<\cdots<x_n\), and \({\mathbf y}_i={\mathbf g}(x_i)\;(i=1,\dots,n)\), to ascertain existence and uniqueness of a function of the form \((\ast)\) with \({\mathbf b}_1,{\mathbf b}_2,\dots{\mathbf b}_n\in{\mathbb R}^m\) with \[ {\mathbf f}(x_i)={\mathbf y}_i,\;i=1,2,\dots,n, \] and to give upper bounds for \(\|{\mathbf f}(u)-\overline{{\mathbf y}}\|\); \(\|\cdot \|\) is the Euclidean norm on \({\mathbb R}^m\) and \(\overline{{\mathbf y}}\) is a given vector from \({\mathbb R}^m\). The layout, after the introductory section 1, is as follows: 2. Preliminary results - proving existence and uniqueness with the aid if generalized Vandermonde determinants - stating the main result 3. Preparatory results - mainly estimates for the values of the generalized Vandermonde determinants 4. Proof of the main result 5. Remarks and examples - discussion and an explicit example