Boundary behavior of the Bergman kernel function on pseudoconvex domains with comparable Levi form. (Q1405293)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Boundary behavior of the Bergman kernel function on pseudoconvex domains with comparable Levi form. |
scientific article |
Statements
Boundary behavior of the Bergman kernel function on pseudoconvex domains with comparable Levi form. (English)
0 references
25 August 2003
0 references
Many authors have studied the problem of determining geometric quantities that are equivalent to the size of the Bergman kernel function on the diagonal near boundary points of smooth, bounded, pseudoconvex domains in~\(\mathbb{C}^{n}\). Here the author treats the case of a boundary point having a neighborhood in which the boundary is of finite type and all the eigenvalues of the Levi form are comparable to each other (that is, there exists a positive constant~\(C\) such that every two eigenvalues \(\lambda\) and~\(\mu\) satisfy the inequality \(C^{-1} \mu(z) \leq \lambda(z) \leq C \mu(z)\) for every boundary point~\(z\) in the neighborhood). The statement of the main theorem requires some notation. Let \(z_{0}\) be a boundary point of finite type~\(m\), let \(U\) be a sufficiently small neighborhood of~\(z_0\), let the eigenvalues of the Levi form be comparable in~\(U\) in the above sense, let \(r\) be a defining function for the domain, let the coordinates be chosen such that the derivative \(\partial r/\partial z_{n}\) is bounded away from zero in~\(U\), let \(L\) be the tangential vector field \((\partial/\partial z_{1}) - (\partial r/\partial z_{n})^{-1} (\partial r/\partial z_{1}) (\partial/\partial z_{n})\) (another canonical vector field would serve as well, since the eigenvalues of the Levi form are comparable), let \(\mathcal{L}_{j,k} \partial \overline{\partial} r(z)\) denote (for positive integers \(j\) and~\(k\)) the derivative of the Levi form given by \({L\mathstrut}^{j-1} {\overline{L}}^{k-1}\partial \overline{\partial}r(z)(L,\overline{L}\,)\), let \(C_l(z)\) denote \(\max\{| \mathcal{L}_{j,k} \partial \overline{\partial} r(z)| : j+k=l\}\), and let \(M(z)\) denote the sum \(\sum_{l=2}^{m} C_{l}(z)^{2(n-1)/l} | r(z)| ^{-2(n-1)/l}\). Then the Bergman kernel function on the diagonal at a point~\(z\) in the intersection of the domain with the neighborhood~\(U\) is comparable to (that is, bounded above and below by positive constants times) the quantity \(M(z)| r(z)| ^{-2}\). In other recent articles, the author has studied on the same class of domains both invariant metrics and derivatives of the Bergman kernel function [J. Math. Kyoto Univ. 42, 337--349 (2002; Zbl 1036.32013); J. Korean Math. Soc. 39, 425--437 (2002; Zbl 1013.32001)].
0 references
Bergman kernel function
0 references
finite type boundary
0 references
eigenvalues of the Levi form
0 references
0 references
0 references
0 references
0 references
0 references