Embedding theorems for Müntz spaces (Q424837)

The paper deals with embeddings of Müntz spaces of \(L^1([0,1])\) in \(L^1(\mu)\). Let us recall that a Müntz subspace \(M_\Lambda^1\) of \(L^1([0,1])\) is the closure (relative to the \(L^1\)-norm) of the subspace spanned by the monomials \(x^{\lambda_n}\), where \(\Lambda=\{\lambda_n\}\) denotes an increasing sequence of positive real numbers. Moreover, it is well known that \(\sum1/\lambda_n<+\infty\) if and only if \(M_\Lambda^1\subsetneq L^1([0,1])\). In the following, this latter condition is always fulfilled. The functions belonging to \(M_\Lambda^1\) are actually real analytic (and extend to analytic function on the unit disc when the \(\lambda_n\)'s are integers). The geometry of Müntz spaces is still rather unknown (see the monograph of \textit{V. I. Gurariy} and \textit{W. Lusky} [Geometry of Müntz spaces and related questions. Lecture Notes in Mathematics 1870. Berlin: Springer (2005; Zbl 1094.46003)] for miscellaneous results on the subject). On the other hand, a large literature exists on embeddings of (more or less) classical spaces of analytic spaces (on the unit disc \({\mathbb D}\) for instance) in \(L^p({\mathbb D},\mu)\) spaces. Spaces like Hardy, Bergman, Dirichlet, \(\dots\) are maybe the most often considered. Very often, conditions involve Carleson measures. Hence, it is natural to wonder under which condition(s) we can embed \(M_\Lambda^1\) in some \(L^1(\mu)\). This work was initiated in the PhD thesis of \textit{Ihab Al Alam} (see [J. Math. Anal. Appl. 358, No. 2, 273--280 (2009; Zbl 1170.47011)], too). In the paper under review, some necessary and some sufficient conditions are given to ensure the boundedness or compactness of such embeddings. The authors consider the maximal function \[ K(u)=\inf\Big\{C\geq0\,\Big|\;\|f\|_{\infty,[0,u]}\leq C\int_{[u,1]}|f(x)|dx\, ,\, \forall\, f\in M_\Lambda^1 \Big\} \] Among miscellaneous interesting results, let us mention the following sufficient criterion. Theorem 2.6. If \(K\in L^1(\mu)\), the embedding \(i_\mu:M^1_\Lambda\longrightarrow L^1(\mu) \) is bounded with norm less than \(\|K\|_{L^1(\mu)}\). Theorem 3.5 expresses the essential norm of \(i_\mu\).