Dirichlet problem of complex Monge-Ampère equation near an isolated KLT singularity

Xin FuarXiv 2021 · DOI · arXiv

Abstract

We solve the Dirichlet Problem of the Monge-Ampère equation near an isolated Kawamata Log Terminal (Klt) singularity. This work generalizes the results of Eyssidieux-Guedj-Zeriahi, where the Monge-Ampère equation was solved on singular varieties without boundary. As a corollary, we construct solutions to the Monge-Ampère equation with isolated singularities on strongly pseudoconvex domains $\Omega$ contained in $\mathbb{C}^n$.

Introduction

In complex geometry, the complex Monge-Ampère (CMA) equation is the fundamental tool used to construct canonical metrics, such as Kähler-Einstein metrics. While the smooth case was famously settled by Yau in the 1970s, modern algebraic geometry frequently encounters "singular" spaces—varieties that possess sharp corners, cusps, or pinches.

A specific class of "mild" singularities known as Kawamata Log Terminal (Klt) singularities is particularly significant within the Minimal Model Program. Previous research, notably by Eyssidieux, Guedj, and Zeriahi (EGZ), demonstrated how to solve the CMA equation on compact singular varieties. However, many local geometric problems—such as understanding the "tangent cone" or the local behavior of a singularity—require solving the Dirichlet problem: finding a solution on a local domain with prescribed values on the boundary. This paper addresses this gap by solving the Dirichlet problem near isolated Klt singularities.

The Challenge of Singularity and Boundary

The difficulty in this setting is twofold. First, one must manage the algebraic degeneracy inherent in the Klt singularity. Second, one must maintain rigorous analytical control over the solution as it approaches the boundary of the domain. Unlike the global case, where the compactness of the manifold provides certain structural advantages, the local Dirichlet problem requires a delicate balance between the singularity's "badness" and the boundary's constraints.

Methodology

The proof utilizes a combination of algebraic geometry and high-order partial differential equation (PDE) estimates. The approach can be broken down into four primary stages:

  1. Log Resolution: The problem is "lifted" from the singular variety $X$ to a smooth variety $Y$ via a birational morphism. This transformation turns the singularity into a set of hypersurfaces (divisors) that are more amenable to analytical treatment.
  2. Perturbation Method: To address the degeneracy caused by the singularity, a perturbed family of equations is introduced: $(\omega + s\theta + \sqrt{-1}\partial\bar{\partial}\phi_s)^n = f_s$. Here, $s$ is a small positive parameter that ensures the equation remains non-degenerate during the approximation process.
  3. Pluripotential Theory: The author adapts Kolodziej’s $C^0$ estimates to the Dirichlet setting. By employing capacities—a method for measuring the "size" of sets in complex analysis—the author proves that the solutions $\phi_s$ remain uniformly bounded as the perturbation parameter $s$ approaches zero.
  4. Barrier Functions and Kodaira Lemma: To achieve higher-order regularity and ensure the metric is smooth away from the singularity, the author utilizes the Kodaira Lemma to construct specific barrier functions. These functions provide the necessary "positivity" to dominate singular terms in the equation.

Key Insights

Klt as the Analytical Sweet Spot

This research reinforces the notion that Klt singularities represent a natural boundary for the solvability of second-order elliptic PDEs. The integrability conditions inherent in Klt singularities allow the right-hand side of the Monge-Ampère equation to remain within $L^p$ spaces, which is the critical requirement for applying Kolodziej’s estimates.

Local vs. Global Transition

The primary innovation lies in transitioning from global results on closed manifolds to local results on domains with boundaries. The author demonstrates that the positivity required to solve the equation can be recovered locally by carefully selecting reference forms and leveraging the geometry of the blowing-up process.

Ricci-Flatness

The solutions constructed are shown to be Ricci-flat ($Ric(g_{KE}) = 0$) on the smooth locus of the domain. This provides a robust local model for how gravity-like curvature behaves near an algebraic pinch point.

Results and Corollaries

Theorem 1.4 (Main Result): For a domain $U$ near an isolated Klt singularity $p$, there exists a unique solution to the Dirichlet problem that is smooth on the regular locus ($U \setminus \{p\}$) and bounded near the singularity.

Corollary 1.1: The author constructs singular solutions on strongly pseudoconvex domains in $\mathbb{C}^n$. These solutions exhibit a specific logarithmic growth rate: $|\phi - s \log(\sum |z_i|^2)| = O(1)$. This effectively prescribes the "strength" of the singularity at the origin, allowing for more precise modeling of singular Kähler metrics.

Limitations and Future Directions

While this work marks a significant advancement, it focuses specifically on isolated singularities. In higher dimensions, singularities often occur along curves or surfaces (non-isolated). Extending these results to non-isolated cases would require a more complex analysis of the exceptional locus. Furthermore, while the solution is bounded and smooth away from the singularity, the exact asymptotic behavior of the metric at the singularity remains an open area of study, particularly in the context of the Donaldson-Sun conjecture.

Summary updated

Abstract

We solve the Dirichlet Problem of Monge-Ampère equation near an isolate Klt singularity, which generalizes the result of Eyssidieux-Guedj-Zeriahi \cite{EGZ}, where the Monge-Ampère equation is solved on singular varieties without boundary. As a corollary, we construct solutions to Monge-Ampère equation with isolated singularity on strongly pseudoconvex domain $Ω$ contained in $\mathbb{C}^n$.
BibTeX
@misc{fu2021dirichlet,
  author = {Xin Fu},
  title = {{Dirichlet problem of complex Monge-Ampère equation near an isolated KLT singularity}},
  year = {2021},
  eprint = {2106.10389},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2106.10389},
}
Figure