On minimal Haar coverings of \(\Bbbk^2\) by three-dimensional spaces (Q2447631): Difference between revisions

Let \(\mathbb K\) be either \(\mathbb R\) or \(\mathbb C\); let \(C_{\mathbb K} (K)\) be the space of continuous \(\mathbb K\)-valued functions on the topological space \(K\); and let \(\mathbb K[x_1,\dots,x_d]\) denote the algebra of polynomials in \(d\) variables, with coefficients in \(\mathbb K\). For \(Z = \{z_1,\dots,z_n\} \subset K\), with the \(z_j\) distinct, and \(H\) an \(n\)-dimensional subspace of \(C(K)\), the interpolation problem \((H,Z)\) is well posed if, for each \(\{c_1,\dots,c_n\}\subset \mathbb K\), there exists an unique \(h \in H\) such that \(h(z_j) = c_j\) , for all \(j\). An \(n\)-dimensional subspace \(H\subset C_{\mathbb K} (K)\) is a Haar space if each nonzero \(h \in H\) has at most \(n-1\) zeros. Then an \(n\)-dimensional subspace, \(H\), has well-posed interpolation on each set of \(n\) distinct points iff \(H\) is a Haar subspace. As is known, the space \(P_{n-1}\subset \mathbb K[x]\), of polynomials of degree at most \(n\), is a Haar space. Indeed, in \(\mathbb C[x], P_{n-1}\) is the unique Haar subspace of dimension \(n\). On the other hand, \textit{J. C. Mairhuber} [Proc. Am. Math. Soc. 7, 609--615 (1956; Zbl 0070.29101)] has shown that, for \(n>1\), no \(n\)-dimensional subspace of \(C_{\mathbb R}(\mathbb{R}^2)\) is Haar; the analogous result, for the complex case has been established by \textit{I. J. Schoenberg} and \textit{C. T. Yang} [Ann. Mat. Pura Appl. (4) 54, 1--12 (1961; Zbl 0099.04602)]; and these results have motivated the authors to define Haar-like coverings of \(C(K)\). In the present article, the authors consider the following question: What is the smallest number, \(\eta_n (\mathbb K^2)\), of \(n\)-dimensional subspaces, \(H_j\) , such that \(\{H_j: j = 1,\dots,\eta_n (\mathbb K^2) \}\) is a Haar covering of \(\mathbb K^2\)? The authors have observed that, for every \(n\), there is a finite Haar covering of \(K\), and \textit{P. Gordan} [Journ. de Math. (5) 6, 141--156 (1900; JFM 31.0108.01)] has demonstrated the existence of finitely many \(n\)-dimensional subspaces of \(\mathbb K[x_1,\dots,x_d]\) that complement every ideal of codimension \(n\). ``Based on various considerations in Algebraic Geometry, Kyungyong Lee proposed a `working conjecture' '': \(\eta_n(\mathbb C^2 ) = n\), and the principal result of the present work shows that \(\eta_3 (\mathbb K^2) = 3\). The article concludes with a section that contains several additional remarks and problems; there is yet work to be done.