Towards a \(p\)-adic Müntz theorem (Q2574789)

The classical Weierstrass theorem states that every \(f \in C[a,b]\) (the space of all continuous functions \([a,b] \rightarrow \mathbb{R}\), \(a,b \in \mathbb{R}\), \(a < b\)) can uniformly be approximated by polynomial functions. It is a natural question to ask which monomials can be deleted in the approximation process. The answer was given by \textit{Ch. H. Müntz} in [Schwarz--Festschr. 303--312 (1914; JFM 45.0633.02)]: A sequence \(n_0 < n_1 < \ldots\) in \(\{ 0,1,2, \ldots \}\) has the property that the linear span of \(x \mapsto x^{n_i}\) (\(i \in \{ 0,1,2, \ldots \}\)) is uniformly dense in \(C[0,1]\) if and only if \(n_0=0\) and \(\sum_{i=1}^{\infty} 1/n_i = \infty\). In the present paper, the author considers the \(p\)-adic case. Let \(p\) be a prime number and let \(\mathbb{Q}_p\) be the field of the \(p\)-adic numbers with the usual \(p\)-adic valuation \(\left| \; . \; \right| \). For an infinite compact subset \(X\) of \(\mathbb{Q}_p\), let \(C(X)\) be the Banach space of all continuous functions \(X \rightarrow \mathbb{Q}_p\), equipped with the maximum norm. The paper is concerned with determining the Müntz sets for \(X\), i.e., the subsets \(S\) of \(\mathbb{N}_0:=\{ 0,1,2, \ldots \}\) having the property that the linear span of \(x \mapsto x^{n}\), \(n \in S\), is uniformly dense in \(C(X)\). The starting point is the well-known \(p\)-adic version of the Weierstrass theorem, proved by \textit{I. Kaplansky} in [Proc. Am. Math. Soc. 1, 356--357 (1950; Zbl 0038.07002)]: Every continuous function \(X \rightarrow \mathbb{Q}_p\) can uniformly be approximated by polynomial functions. Then, in a first step, the author describes the Müntz sets for \(\mathbb{B}_p:= \{ x \in \mathbb{Q}_p : \left| 1-x \right| <1 \}\): A subset of \(\mathbb{N}_0\) is a Müntz set for \(\mathbb{B}_p\) if and only if it is infinite. In a second step, he describes the Müntz sets for \(\mathbb{T}_p:= \{ x \in \mathbb{Q}_p : \left| x \right| = 1 \}\): If \(p \neq 2\), a subset \(S\) of \(\mathbb{N}_0\) is a Müntz set for \(\mathbb{T}_p\) if and only if, for each \(j \in \{ 0,1, \ldots, p-2 \}\), \(S \cap \{ n \in \mathbb{N}_0 : n \equiv j\) (mod\((p-1)) \}\) is infinite. If \(p=2\), a subset \(S\) of \(\mathbb{N}_0\) is a Müntz set for \(\mathbb{T}_p\) if and only if \(S\) contains infinitely many even numbers and infinitely many odd numbers. Finally, several sufficient conditions are derived for a set to be a Müntz set for \(\mathbb{Z}_p:= \{ x \in \mathbb{Q}_p : \left| x \right| \leq 1 \}\); as an application, some interesting examples of such Müntz sets are obtained. To achieve his purpose, the author carries out nice reasonings consisting of clever combinations of tools and ideas coming from different parts of \(p\)-adic analysis, such as calculus and functional analysis. The first section of the paper contains the background information needed to follow his argumentation.