Partial Regularity of Harmonic Maps From Alexandrov Spaces

Huabin Ge, Wenshuai Jiang, Hui-Chun ZhangInternational Mathematics Research Notices 2019 · arXiv

TL;DR

This paper proves that continuous harmonic maps from Alexandrov spaces to smooth Riemannian manifolds are Lipschitz continuous, not just Hölder continuous as previously known. This resolves a long-standing conjecture and extends existing regularity results for harmonic maps to a more general class of singular spaces.

What Problem Does This Paper Address?

Harmonic maps are fundamental objects in geometry and physics, analogous to harmonic functions in Euclidean space. For maps between smooth manifolds, their regularity (how smooth they are) has been extensively studied. However, understanding harmonic maps into singular spaces, specifically Alexandrov spaces (which are metric spaces with a lower bound on curvature but can have singularities), is more challenging.

F. H. Lin had previously shown that harmonic maps from Alexandrov spaces are partially Hölder continuous, meaning they are Hölder continuous everywhere except possibly on a set of small Hausdorff dimension. This left open the question of whether this Hölder continuity could be improved to the stronger Lipschitz continuity. This paper tackles this problem, aiming to prove Lipschitz regularity for all continuous harmonic maps from Alexandrov spaces to smooth Riemannian manifolds.

Methods

The authors build upon the work of Huang-Wang [26] and extend their techniques for smooth domains to the setting of Alexandrov spaces. The core of their approach involves:

  • Leveraging Alexandrov Space Properties: They make crucial use of the specific geometric properties of Alexandrov spaces, particularly those related to their local structure, volume growth (Bishop-Gromov inequality), and the existence of a smooth structure on a large subset of the space.
  • Heat Flow and Poisson Equations: The proof utilizes estimates derived from heat kernels and solutions to Poisson equations on Alexandrov spaces. This is a key adaptation, as direct use of Green functions (common in Euclidean settings) is not straightforward in these singular spaces.
  • Morrey-Type Estimates: A central technical tool is the derivation of a Morrey-type decay estimate for the gradient of the harmonic map. This estimate relates the $L^2$ norm of the gradient in a small ball to the size of the ball and the $L^\infty$ norm of the map.
  • Local Lipschitz Estimates: The paper establishes local Lipschitz estimates for solutions to Poisson equations with $L^1$ data, which is crucial for controlling the gradient.
  • Iterative Inequalities: The proof involves applying inequalities iteratively to shrink the domain and improve the regularity estimates, ultimately leading to the Lipschitz continuity.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Lipschitz Regularity is Achievable: The most significant insight is that the Lipschitz regularity, which was previously only conjectured, holds true for harmonic maps from Alexandrov spaces. This is a substantial improvement over the known partial Hölder continuity.
  • Bridging Smooth and Singular Geometry: The paper successfully bridges the gap between the well-understood regularity theory of harmonic maps on smooth manifolds and the more complex setting of singular Alexandrov spaces. This is achieved by carefully adapting techniques from smooth geometry to the Alexandrov space framework.
  • Importance of Alexandrov Space Structure: The proof highlights that the specific geometric structure of Alexandrov spaces (beyond just being metric spaces with bounded curvature) is essential for obtaining Lipschitz regularity. Simply having a smooth target manifold and a lower bound on curvature is not sufficient; deeper properties of the domain space are leveraged.
  • Adaptation of Analytical Tools: The paper demonstrates the power of adapting analytical tools like heat kernels and Poisson equation estimates to the context of Alexandrov spaces, overcoming the challenges posed by their singularities.

Key Results

The main quantitative result is the proof of Theorem 1.3: Any continuous harmonic map $u$ from a bounded domain $\Omega$ in an $n$-dimensional Alexandrov space with curvature bounded below to a compact smooth Riemannian manifold $N$ is locally Lipschitz continuous in $\Omega$. This is achieved by showing that if the oscillation of $u$ in a small ball is sufficiently small, then $u$ is Lipschitz continuous in that ball.

As a direct consequence, Theorem 1.4 states that $u$ is locally Lipschitz continuous away from a relatively closed subset of Hausdorff dimension $\le n-3$. This resolves Conjecture 1.2 posed by F. H. Lin.

Limitations and Future Work

  • Energy Minimizers vs. Continuous Maps: While the paper proves Lipschitz continuity for any continuous harmonic map, F. H. Lin's earlier work focused on energy minimizing maps. The current result is more general, but the techniques might be further refined for energy minimizers.
  • Smooth Target Manifold: The target manifold $N$ is assumed to be smooth. Extending these results to targets with singularities would be a significant next step.
  • Generalization of Alexandrov Spaces: The paper focuses on finite-dimensional Alexandrov spaces. Investigating harmonic maps from infinite-dimensional Alexandrov spaces or other types of singular spaces could be a direction for future research.
  • Higher Regularity: The paper establishes Lipschitz continuity. Investigating whether $C^1$ or even $C^\infty$ regularity can be achieved under certain conditions remains an open question.

Summary updated

Abstract

Abstract In this paper, we prove the Lipschitz regularity of continuous harmonic maps from a finite-dimensional Alexandrov space to a compact smooth Riemannian manifold. This solves a conjecture of F. H. Lin in [38]. The proof extends the argument of Huang-Wang [28].
BibTeX
@inproceedings{ge2019partial,
  author = {Huabin Ge and Wenshuai Jiang and Hui-Chun Zhang},
  title = {{Partial Regularity of Harmonic Maps From Alexandrov Spaces}},
  booktitle = {International Mathematics Research Notices},
  year = {2019},
}
Figure