Nonuniqueness of the solutions to the asymptotic Plateau problem

Barış CoşkunuzerarXiv 2006 · DOI · arXiv

TL;DR

This paper tackles the question of whether every boundary condition for the asymptotic Plateau problem in hyperbolic space leads to a unique area-minimizing surface. The authors demonstrate that this is not the case, proving the existence of boundary curves that admit multiple distinct area-minimizing surfaces. This finding is significant as it challenges a potential assumption of uniqueness in this geometric problem.

What Problem Does This Paper Address?

The asymptotic Plateau problem in hyperbolic space concerns finding a complete, area-minimizing surface within hyperbolic space ($H^{n+1}$) whose boundary at infinity matches a given closed submanifold (the "asymptotic boundary") on the sphere at infinity ($S^n(H^{n+1})$). While Michael Anderson proved in 1982 that at least one such surface always exists for any given asymptotic boundary, the question of how many such surfaces exist remained open. Previous work had established uniqueness for specific types of boundaries (e.g., those bounding mean-convex domains). This paper directly addresses the possibility of non-uniqueness, seeking to determine if there are boundary conditions that allow for multiple distinct area-minimizing solutions.

Methods

The core of the paper's method involves constructing a specific type of asymptotic boundary that forces non-uniqueness. The authors leverage existing results, particularly those concerning minimal surfaces (a subset of area-minimizing surfaces) in hyperbolic space.

  1. Leveraging Minimal Surface Non-Uniqueness: They start by referencing Theorem 2.3, which states that there exist simple closed curves in $S^2(H^3)$ that bound infinitely many complete minimal surfaces in $H^3$.
  2. Constructing a Folitation: The key idea is to create a "foliation" of the sphere at infinity ($S^n(H^{n+1})$) with a family of (n-1)-spheres, $\{ \Gamma_t \}$. One of these spheres, $\Gamma_0$, is chosen to be an asymptotic boundary that is known to admit multiple minimal surfaces.
  3. The Uniqueness Assumption and Contradiction: The authors then assume, for the sake of contradiction, that every asymptotic boundary in this foliation admits a unique absolutely area-minimizing hypersurface.
  4. The Folitation of Hyperbolic Space: Under this uniqueness assumption, the corresponding unique area-minimizing hypersurfaces $\{ \Sigma_t \}$ would then form a foliation of the entire hyperbolic space $H^{n+1}$.
  5. The Contradiction: However, since $\Gamma_0$ is known to bound multiple minimal surfaces, at least one of these minimal surfaces must intersect the foliation $\{ \Sigma_t \}$ in a way that violates the maximum principle for minimal hypersurfaces. This contradiction proves that the initial assumption of uniqueness must be false for at least one boundary in the foliation.

The proof is then extended to higher dimensions by constructing higher-dimensional minimal hypersurfaces from the 2D minimal surfaces.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Non-uniqueness is possible: The most crucial insight is that the asymptotic Plateau problem does not always have a unique solution. This means that for some "boundary conditions" at infinity, there can be multiple distinct area-minimizing surfaces that satisfy them.
  • Minimal surfaces as a stepping stone: The proof cleverly uses the known non-uniqueness for minimal surfaces to construct non-uniqueness for absolutely area-minimizing surfaces. This highlights the close relationship between these two classes of surfaces.
  • Foliations as a proof technique: The construction of a foliation of the sphere at infinity and the subsequent argument about the foliation of hyperbolic space is a powerful geometric technique to establish the existence of non-unique solutions.
  • The role of the maximum principle: The maximum principle for minimal hypersurfaces is essential for deriving the contradiction that proves the existence of non-unique solutions.

Key Results

  • Theorem 3.2: For any dimension $n \geq 2$, the paper proves the existence of codimension-1 closed submanifolds $\Gamma$ in $S^n(H^{n+1})$ such that there are more than one absolutely area-minimizing rectifiable n-currents $\Sigma$ in $H^{n+1}$ with $\partial_\infty \Sigma = \Gamma$.
  • Extension to Area Minimizing Planes (Remark 3.1): The paper also suggests that similar ideas can be used to show non-uniqueness for area-minimizing planes in $H^3$, demonstrating that a simple closed curve in $S^2(H^3)$ can bound more than one area-minimizing plane.

Limitations and Future Work

  • Existence, not examples: The paper proves the existence of such non-unique boundary conditions but does not provide explicit, concrete examples of these submanifolds. The authors note that finding such explicit examples is challenging due to the noncompact nature of the problem.
  • Quantitative data: The authors mention that in compact settings, quantitative data can help find examples, but these techniques don't directly apply to the asymptotic case.
  • Topological properties: The paper suggests that the main theorem could also be proven by examining the topological properties of foliations of disks, potentially using Euler characteristic arguments. This points to alternative avenues for understanding and potentially constructing such examples.
  • Generalization to other spaces: While this paper focuses on hyperbolic space, future work could explore similar questions in other geometric settings.

Summary updated

Abstract

Because of the relevance of the results, this paper is merged into the paper titled "On the Number of Solutions to Asymptotic Plateau Problem" (arXiv:math.DG/0505593) as a new section.
BibTeX
@misc{coskunuzer2006nonuniqueness,
  author = {Barış Coşkunuzer},
  title = {{Nonuniqueness of the solutions to the asymptotic Plateau problem}},
  year = {2006},
  eprint = {math/0603635},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/math/0603635},
}
Figure