An equivalence theorem of a class of Minkowski norms and its applications

Huitao Feng, Yuhua Han, Ming LiarXiv 2018 · DOI · arXiv

Abstract

In this paper, the Cartan tensors of the $(\alpha,\beta)$-norms are investigated in detail to establish a new equivalence theorem. As a primary application in Finsler geometry, it is demonstrated that general $(\alpha,\beta)$-metrics on smooth manifolds of dimension $n \geq 4$ with vanishing Landsberg curvatures must be Berwald manifolds. This result effectively rules out the existence of "unicorns"—Landsberg manifolds that are not Berwald—within this significant class of geometric spaces.

Introduction: The Search for Mathematical Unicorns

In the field of Finsler geometry, which generalizes Riemannian geometry by allowing the distance metric to depend on both position and direction, two specific types of manifolds have long been the subject of intense study: Berwald manifolds and Landsberg manifolds. By definition, every Berwald manifold is a Landsberg manifold. However, the converse—whether every Landsberg manifold is necessarily Berwald—has remained one of the most elusive questions in the field.

In 2005, the geometer David Bao famously dubbed the potential existence of a Landsberg manifold that is not Berwald as the "Unicorn Problem." This paper addresses this mystery for a widely-studied family of metrics known as general $(\alpha, \beta)$-metrics, constructed using a Riemannian metric $\alpha$ and a 1-form $\beta$.

Methodology and the Equivalence Theorem

The authors approach the problem by treating the "indicatrix" (the unit sphere in the tangent space) as a hypersurface within the framework of centroaffine differential geometry. By studying the induced affine metric and the cubic form of these hypersurfaces, they develop a new equivalence theorem for Minkowski norms.

The core of their proof rests on the discovery that if two Minkowski norms share certain geometric invariants—specifically the fundamental tensor $\bar{g}$ and the Cartan form $\eta$—they must be essentially identical. To bridge the gap between local Minkowski norms and global manifold properties, the authors utilize nonlinear parallel transport. In Landsberg spaces, parallel transport preserves the Finsler metric; the authors prove that for the $(\alpha, \beta)$ class, this transport must be linear, which is the defining characteristic of a Berwald space.

Furthermore, the research introduces a "characteristic function" $q$ for $(\alpha, \beta)$-norms. This function serves as a classification tool and a critical component in the analytical proof that Landsberg curvature vanishing implies the Berwald property.

Key Results

The $n \geq 4$ Threshold

A significant finding of the paper is the dimension threshold required for the proof. The authors establish that for dimensions $n \geq 4$, any general $(\alpha, \beta)$-metric with vanishing Landsberg curvature is Berwald. This reliance on higher dimensions stems from the application of Schur’s Lemma to the indicatrix to demonstrate that certain geometric forms have constant norms—a technique that is mathematically invalid in dimensions 2 and 3.

Rigidity of Minkowski Spaces

The paper also establishes a new rigidity result (Theorem 5). It shows that the metric and the Cartan form together uniquely determine the Minkowski norm up to a sign. This provides a fundamental constraint on the geometry of these spaces and generalizes previous work that required more restrictive conditions, such as the vanishing of the mean Berwald curvature ($E=0$).

Analyticity and Global Extension

By leveraging the analyticity of the functions defining these metrics, the authors are able to extend results from specific "non-Euclidean" regions of the tangent bundle to the entire space. This ensures that if the Berwald property holds in a specific region, it must hold across the entire manifold.

Conclusion and Future Directions

While this paper provides a definitive answer for general $(\alpha, \beta)$-metrics in dimensions $n \geq 4$, the "Unicorn Problem" is not yet fully extinct. The authors explicitly note that their current methods do not apply to lower dimensions ($n=2, 3$), where unicorns might still theoretically roam. Additionally, while $(\alpha, \beta)$-metrics are a broad and important class, they do not encompass all possible Finsler metrics.

Future research may focus on investigating the characteristic function $q$ further, particularly its potential relationship to Lagrangian submanifolds, and extending these rigidity results to other classes of Finsler metrics to finally determine if the "unicorn" is a mathematical reality or a geometric myth.

Summary updated

Abstract

In this paper, the Cartan tensors of the $(α,β)$-norms are investigated in details. Then an equivalence theorem of $(α,β)$-norms is proved. As a consequence in Finsler geometry, general $(α,β)$-metrics on smooth manifolds of dimension $n\geq4$ with vanishing Landsberg curvatures must be Berwald manifolds.
BibTeX
@misc{feng2018equivalence,
  author = {Huitao Feng and Yuhua Han and Ming Li},
  title = {{An equivalence theorem of a class of Minkowski norms and its applications}},
  year = {2018},
  eprint = {1812.11938},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1812.11938},
}
Figure