Sharp explicit lower bounds of heat kernels (Q1381575): Difference between revisions
From MaRDI portal
Latest revision as of 10:35, 28 May 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Sharp explicit lower bounds of heat kernels |
scientific article |
Statements
Sharp explicit lower bounds of heat kernels (English)
0 references
4 August 1999
0 references
Let \(M\) be a complete Riemannian manifold of dimension \(d\) and let \(L={1\over 2}(\Delta+Z)\) where \(Z\) is a \(C^1\)-vector field. Denote by \(\rho_t(x,y)\) the heat kernel of \(L\). The author uses a probabilistic approach called ``logarithmic transformation'' to prove explicit lower bounds for \(\rho_t\). The advantage of the probabilistic approach by \textit{W. H. Fleming} [Lect. Notes Contr. Inf. Sci. 42, 131-141 (1982; Zbl 0502.93076)] is that it can be applied also in the case of unbounded curvature. A major assumption made in this paper is that the Ricci curvature of \(M\) can be bounded from below by a function depending quadratically on the distance function \(\rho_y(x)\), i.e., \(\text{Ric}(X,X)\geq-(k_1+k_2\rho_y(x))^2(d-1) | X|^2, X\in T_xM\), where \(k_i, i=1,2\) are positive constants depending on \(y\) and \(y\in M\) is a fixed point. Further he assumes that \(b_2:=\sup\{\langle\nabla_UZ,U\rangle:U\in TM, | U|=1\}\) is bounded and defines \(b_1:=| Z(y)|\) and \(\beta_i:=\max\{0,(d-1)k_i+ b_i\}\). The main theorem then states that for any \(\sigma>0\) one has \[ \rho_t(x,y)\geq (2\pi t)^{-d/2}\exp\{-f_1(\sigma,t)\rho^2 (x,y)-f_2(\sigma, t)\}, \quad t>0,\;x,y\in M, \] where \(f_i, i=1,2\) are two positive functions given in the paper and depending on terms involving \(\sqrt{t}\) up to fifth order and on \(\sigma,\beta_1, \beta_2\) and \(d\).
0 references
heat kernel
0 references
lower bounds
0 references
Ricci curvature
0 references
distance function
0 references
logarithmic transformation
0 references
0 references