Approximation by Chebyshevian Bernstein operators versus convergence of dimension elevation (Q295816)

The main result of the paper involves a nested sequence \[ \mathbb{E}_{1}\subset \mathbb{E}_{2}\subset\dots\subset \mathbb{E}_{n}\subset\dots \] where \(\mathbb{E}_{1}\) contains constants and each \(\mathbb{E}_{n}\) is an \( (n+1)\)-dimensional extended Chebyshevian space included in \(C^{\infty }\left( [a,b]\right) \). Moreover, consider a Bernstein basis \(\left( B_{n,0},B_{n,1},\dots,B_{n,n}\right) \) of \(\mathbb{E}_{n}\). In this way one obtains an approximation process called by the authors Bernstein approximation process. Another approximation process associated to the above described sequence of extended Chebyshevian spaces,is the so called infinite dimension elevation process, which involves the Bézier points. The main result of the paper is Theorem 4.3, where it is proved that the these two approximation methods are equivalent from the convergence point of view. Moreover, other two equivalent characterizations are proposed and one of them is actually a Korovkin type characterization of uniform convergence. In order to prove the theorem, many interesting auxiliary results are proved in a separate section. We mention here the definition of a uniformly dense grid and many uniform convergence properties associated to these types of grids. The results are further generalized in Theorem 5.3, where truncated nested sequences are involved. Lastly, the results are applied considering concrete examples such as linear differential operators with constant coefficients.