Abstract:We present a method of obtaining a lower bound estimate of the curvatures of the Bergman metric without using the regularity of the kernel function on the boundary. As an application, we prove the existence of an uniform lower bound of the bisectional curvatures of the Bergman metric of a smooth bounded pseudoconvex domain near the boundary with constant Levi rank.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is whether there is a uniform lower bound estimate for the bisectional curvature of the Bergman metric in a smooth bounded pseudoconvex domain when the boundary has a constant Levi rank. Specifically, the author hopes to obtain a lower bound estimate of the Bergman metric curvature without using the regularity of the kernel function on the boundary. The importance of this problem lies in that the curvature of the Bergman metric is an important invariant in complex differential geometry and plays a crucial role in understanding the geometric properties of the domain.
The main contribution of the paper is to provide a method to prove the existence of such a lower bound without relying on the boundary regularity of the kernel function. This improves the previous research results to a certain extent, especially McNeal's results, because the holomorphic sectional curvature usually cannot control the bisectional curvature and Ricci curvature. By introducing Hörmander's $L^2$ estimate and constructing plurisubharmonic functions with large Hessians, the author successfully proves the main theorem, that is, for a pseudoconvex domain with a smooth boundary, if the Levi form has a constant rank $l$ in a certain neighborhood near $z_0$, then there exists a neighborhood $W$ of $z_0$ such that all bisectional curvatures of the Bergman metric of $\Omega$ are uniformly bounded from below by a negative constant in $W$. In particular, the holomorphic sectional curvature and Ricci curvature are also uniformly bounded.
In addition, the author also shows that this method can be applied to other types of pseudoconvex domains, such as finite - type domains and convex domains, further expanding the application range of the results.