Concentration et randomisation universelle de sous-espaces propres

TL;DR

This paper develops a universal theory of multidimensional randomization in Lebesgue spaces using inequalities, and explores conditions for almost sure convergence of eigenfunction combinations on manifolds, with various applications.

Contribution

It introduces a new method for analyzing eigenfunction convergence based on concentration properties and extends probabilistic techniques to manifold settings.

Findings

Multidimensional randomization is universal in $L^p$ spaces.

Established conditions for almost sure convergence of eigenfunction sums.

Applied results to Sobolev embeddings and nonlinear wave equations.

Abstract

We develop a theory of multidimensional randomization in Lebesgue spaces $L^p$ with the aid of Kahane-Khintchine-Marcus-Pisier inequalities. More precisely, we obtain a result in the spirit of Maurey-Pisier's theorem which involves random matrices and proves that the multidimensional randomization is universal in $L^p$. Then, we deal with the question of studying necessary and sufficient conditions to get the almost sure convergence in $L^p$ of random linear combinations of eigenfunctions in a compact Riemannian manifold (the famous Paley-Zygmund theorem gives the answer for tori). We introduce a new method which solves the problem if the considered eigenfunctions have some concentration property. We give several applications like the almost sure $L^p$ convergence for compact manifolds, spherical harmonics and the harmonic oscillator. We also get probabilistic Sobolev embeddings in the sense of Burq and Lebeau and we obtain global solutions for the supercritical cubic wave equation on a boundaryless compact $3$-manifold.