Positive curvature property for sub-Laplacian on nilpotent Lie group of rank two (Q380032)

From MaRDI portal
Revision as of 03:08, 30 January 2024 by Import240129110155 (talk | contribs) (Added link to MaRDI item.)
scientific article
Language Label Description Also known as
English
Positive curvature property for sub-Laplacian on nilpotent Lie group of rank two
scientific article

    Statements

    Positive curvature property for sub-Laplacian on nilpotent Lie group of rank two (English)
    0 references
    0 references
    11 November 2013
    0 references
    The author aims at obtaining a proof of the following gradient estimate: \[ |\nabla\log p_t(0,g)|\leq C\,d(0,g)/t, \] relative to the heat kernel \(p_t\) of the nilpotent Lie group \(G\) of rank with \(n\) generators (which extend the Heisenberg group). Here, \(d(0,g)\) denotes the Carnot-Carathéodory (subelliptic) distance the origin \(0\) and \(g\in G\). To proceed, the author first notices that a method, as developed by Boudouin, Bonnefart and H. Q. Li, reduces the question to the obtension of a positivity criterion on the \(\Gamma_2\) Bakry-Émery curvature. Thus the main purpose of this article is to obtain such a criterion, actually in some weak but sufficient sense. A convenient realization of \(G\) is obtained by considering independent real Brownian motions \(B_1,\dots, B_n\), together with their \(n(n-1)/2\) pairwise Lévy areas \(A_{jk}\). Its Lie algebra is realized by endowing \(\mathbb{R}^{n(n+1)/2}\) with coordinates \((x_i)_{1\leq i\leq n}\) and \((y_{jk})_{1\leq j< k\leq n}\), and setting \[ Y_{jk}:= {\partial\over\partial y_{jk}},\quad X_i:= {\partial\over\partial x_i}+{1\over 2} \Biggl(\sum^{i-1}_{\ell= 1} x_\ell Y_{\ell i}- \sum^n_{\ell= i+1} x_\ell Y_{i\ell}\Biggr). \] Then the generator of the diffusion process \(((B_i); (A_{jk}))\) is the subelliptic Laplace operator \(\sum^n_{i=1} X^2_i\), from which the caré du champ and the Bakry-Émery curvature \(\Gamma_2\) are defined as usual. Using this realization, the heat semigroup kernel \(p_t\) has a somehow explicit, but involed, expression. The leading idea of the author is to restrict to well-chosen ``radial'' directions, along which careful computations will exhibit a positive restricted \(\Gamma_2\) operator, and which will prove to be sufficient to derive the wanted estimates; basically, once \(p_t(0,\cdot)\) itself is a radial function. In particular, a logarithmic Sobolev inequality is proved to hold for radial functions.
    0 references
    \(\Gamma _{2}\) curvature
    0 references
    heat kernel
    0 references
    gradient estimates
    0 references
    sub-Laplacian
    0 references
    nilpotent Lie groups
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references