Weierstrass type approximation by weighted polynomials in \(\mathbb{R}^d\) (Q2325027)

Weierstrass type approximation by weighted polynomials in \(\mathbb{R}^{d}\) is the subject of the paper. The first main result provides a Weierstrass type result asserting that uniform approximation on \(\mathbb{R}^{d}\) by weighted polynomials \(w^{-2n}p_{2n}\) is possible when the admissible weight \(w\) is convex and, if \(d>1\) is also piecewise \(C^{1}\) (\(p_{n}\) are multivariate polynomials of \(d\) variables of degree at most \(n\)). The similar problem of multivariate polynomial approximation with varying weights is also considered with non convex weights of the form \(w_{\alpha}(\mathbf{x}) := (1+\vert x_1 \vert^{\alpha} + \ldots + \vert x_d \vert^{\alpha})^{\frac{1}{\alpha}},\) where \(\mathbf{x} = (x_1,\ldots,x_d) \in \mathbb{R}^{d}\) and \(0<\alpha <1.\) In this case the necessary condition for weighted polynomial approximation for a given continuous function is that the function vanishes on a certain exceptional set consisting of all coordinate hyperplanes and \(\infty.\) Moreover, this condition is also sufficient in case of rational \(\alpha.\)