scientific article; zbMATH DE number 2196452

From MaRDI portal
Publication:5465296





Finite-dimensional projections whose kernels are polynomial ideals of finite codimension are studied. Let \(\mathbb{C}[x,y]\) stand for the ring of polynomials of two variables with complex coefficients and let \(P\) be a projection from \(\mathbb{C}[x,y]\) onto a finite-dimensional subspace: (1) \(P\) is called an ideal projection iff \(P(f \cdot g)=P(f \cdot Pg)\) for all \(f,g \in \mathbb{C}[x,y]\); (2) \(P\) is called an interpolating projection if there exists a set \(\Delta \subset \mathbb{C}\) such that \(Pf=0\) iff \(f(u)=0\) for all \(u \in \Delta\); (3) \(P\) is called a lim-interpolating projection if there exists a sequence \(P_{k}\) of interpolating projections such that \(P_{k} f \rightarrow Pf\) for every \(f \in \mathbb{C}[x,y]\). If \(P\) is an interpolating projection, then \(P\) is an ideal projection. Hence every lim-interpolating projection is ideal. It is well-known that, for polynomials of one complex variable, the converse is also true. Carl de Boor conjectured that every ideal projection is a lim-interpolating projection. In this paper, the author gives a complete description of ideal projections from \(\mathbb{C}[x,y]\) onto \(\mathbb{C}_{1}[x,y]\), the linear span of \(1,x,y\), and uses it to verify the above-mentioned conjecture for this special case as well as for all three dimensional projections in \(\mathbb{C}[x,y]\). For the entire collection see [Zbl 1061.41001].











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 Q5465296)