The Gaussian limit for high-dimensional spherical means (Q1634599)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The Gaussian limit for high-dimensional spherical means
scientific article

    Statements

    The Gaussian limit for high-dimensional spherical means (English)
    0 references
    0 references
    0 references
    18 December 2018
    0 references
    The paper is devoted to the proof of the following result (Theorem~5.1 in the paper): \par Theorem. Let $H$ be a separable Hilbert space and $B$ the closure of $H$ with respect to a measurable norm. Let $A$ be a closed affine subspace of $H$ of finite codimension. Let $f$ be a bounded Borel function on a finite-dimensional nonzero subspace $V$ of $H$ such that the orthogonal projection $P^H_V : H\to V$ maps $A$ onto $V$. Suppose $Z_1, Z_2,\ldots$ are finite-dimensional subspaces of $H$, and \[ V\subset Z_1\subset Z_2\subset\ldots\subset H \] with $\bigcup_{N\geq1}Z_N$ being dense in $H$. Then \[ \lim\limits_{N\to\infty}\int\limits_{A\cap S_{Z_N}}f_{Z_N} d\overline{\sigma}=\int\limits_H f_B d\mu_A, \] where $\overline{\sigma}$ is the unit-mass uniform surface measure on the ``circle'' $A\cap S_{Z_N}$ in which $A$ intersects the sphere $S_{Z_N}$ of radius $\sqrt{\text{dim }Z_N}$ and center $0$ in the subspace $Z_N$; functions $f_{Z_N}$ and $f_B$ are given via $f_{Z_N}:=f\circ P^{Z_N}_V$, $f_B:=f\circ P^B_V$ (with corresponding orthogonal projections $P^{Z_N}_V : Z_N\to V$ and $P^{B}_V : B\to V$) and $\mu_A$ is the Gaussian measure on $B$ whose Fourier transform is given for all $h\in H$ by \[\exp\left(i\langle p_A,h\rangle-\frac12\|P_0h\|^2 \right) \] with $p_A$ being the point of $A$ closest to $0$ and $P_0 : H\to H$ being the orthogonal projection onto the subspace $A-p_A$. \par The paper is the third in a series, following [Zbl 1270.44003] and [Zbl 1351.44002]. However, the paper is self-contained and can be read independently.
    0 references
    Gaussian Radon transform
    0 references
    spheres
    0 references
    large-$N$ limit
    0 references
    Gaussian measure
    0 references

    Identifiers