Random-weighted Sobolev inequalities on \(\mathbb{R}^d\) and application to Hermite functions (Q2256023): Difference between revisions

Recently \textit{N. Burq} and \textit{G. Lebeau} [Ann. Sci. Éc. Norm. Supér. (4) 46, No. 6, 917--962 (2013; Zbl 1296.46031)] developed a randomization method based on the Laplace operator on a \textit{compact} Riemannian manifold, and showed that almost surely (with respect to the ``natural'' probability measure), a function enjoys better Sobolev estimates than expected. The main object of the paper under review is to extend this randomization method to the case of \(\mathbb{R}^d\) and Sobolev spaces based on the harmonic oscillator in \(L^2(\mathbb{R}^d)\). The main difference and main difficulties are non-compactness of configuration space. The authors ``construct measures, which satisfies the concentration of measure property, on the support of which the optimal-weighted Sobolev estimates on \(\mathbb{R}^d\) are proved. This construction relies on accurate estimates on the spectral function in a non-compact configuration space. As an application, it is shown that there exists a basis of Hermite functions with good decay properties in \(L^\infty(\mathbb{R}^d)\), when \(d \geq 2\).'' The Burq-Lebeau construction of the measures relied on Gaussian random variables, while in the present paper general random variable which satisfy concentration of measure estimates are under consideration. However, the optimal estimates are obtained in the case of the Gaussians only.