The Bohnenblust-Hille inequality for homogeneous polynomials is hypercontractive (Q640791): Difference between revisions

The main goal of this nice and important paper is to obtain that the Bohnenblust-Hille inequality is hypercontractive. The Bohnenblust-Hille inequality says that the \(\ell^{\frac{2m}{m+1}}\)-norm of the coefficients of an \(m\)-homogeneous polynomial on \(\mathbb{C}^n\) is bounded above by the supremum norm of this polynomial times a constant which does not depend on \(n\). The authors show in fact that this constant can be taken to be \(C^m\) for some \(C>1\). This result considerably improves the known bounds, see [\textit{L. A. Harris}, Analyse fonct. Appl., C. r. Colloq. d'Analyse, Rio de Janeiro 1972, 145--163 (1975; Zbl 0315.46040); \textit{H. Queffélec}, J. Anal. 3, 43--60 (1995; Zbl 0881.11068)] and leads to some interesting consequences. For instance, the authors obtain an asymptotically optimal estimate for the \(n\)-dimensional Bohr radius, which was introduced and studied by \textit{H. P. Boas} and \textit{D. Khavinson} [Proc. Am. Math. Soc. 125, No. 10, 2975--2979 (1997; Zbl 0888.32001)], as well as a refined version of a striking theorem of \textit{S. V. Konyagin} and \textit{H. Queffélec} [Real Anal. Exch. 27, No. 1, 155--175 (2002; Zbl 1026.42011)] on Dirichlet polynomials that was recently sharpened by \textit{R. de la Bretèche} [Acta Arith. 134, No. 2, 141--148 (2008; Zbl 1302.11068)]. In the same spirit of the latter, the authors give the following stronger version of a result of \textit{R. Balasubramanian, B. Calado} and \textit{H. Queffélec} [Stud. Math. 175, No. 3, 285--304 (2006; Zbl 1110.30001)]: they prove that \(1/\sqrt{2}\) is the supremum of the set of real numbers \(c\) such that \[ \sum_{n\geq 1}| a_n| n^{-\frac{1}{2}}\text{exp}\Big\{c\sqrt{\log n\log \log n}\Big\}<\infty \] for every ordinary Dirichlet series \(f(s)=\sum_{n\geq 1}a_nn^{-s}\) for which \(\sup\{| f(\sigma+it)| :\;\sigma>0\}<\infty.\)