Length spectrum of large genus random metric maps

Simon Barazer, Alessandro Giacchetto, Mingkun Liu

Abstract

We study the length of short cycles on uniformly random metric maps (also known as ribbon graphs) of large genus using a Teichmüller theory approach. We establish that, as the genus tends to infinity, the length spectrum converges to a Poisson point process with an explicit intensity. This result extends the work of Janson and Louf to the multi-faced case.

Introduction: Geometry in the Large Genus Limit

A ribbon graph is a combinatorial object where vertices possess a cyclic ordering of incident edges, allowing it to be embedded as a thickened surface with a specific genus (number of holes) and boundary components (faces). When positive lengths are assigned to each edge, it becomes a metric ribbon graph. These structures are central to combinatorics and 2-dimensional quantum gravity.

A fundamental question in this field is the behavior of these surfaces as their complexity grows—specifically, the "large genus limit" where the number of holes $g$ approaches infinity. While previous research by Janson and Louf characterized the length spectrum (the set of lengths of short loops) for one-faced (unicellular) maps using combinatorial methods, these techniques do not generalize to multi-faced graphs. This paper provides a universal framework using geometric analysis.

Methodology: A Teichmüller Theory Approach

To overcome the limitations of pure combinatorics, the authors employ tools from hyperbolic geometry and Teichmüller theory:

  1. Combinatorial Moduli Spaces: The study focuses on the space of all metric ribbon graphs of fixed genus $g$ and face count $n$, weighted by the Kontsevich symplectic form.
  2. Integration Formulas: The authors utilize a combinatorial version of Mirzakhani’s integration formula to determine the expected number of cycles of specific lengths.
  3. Asymptotic Analysis: To evaluate the limit as $g \to \infty$, they apply Aggarwal’s results on Witten-Kontsevich intersection numbers, using the saddle-point method to solve the resulting complex integrals.
  4. Method of Moments: This probabilistic technique is used to prove that the distribution of cycle lengths converges to a Poisson point process.

Key Insights: Cycles vs. Curves

A striking finding is the "spooky" parallel between discrete metric graphs and continuous hyperbolic surfaces. The statistical distribution of short cycles on these graphs matches the distribution of short closed geodesics on random hyperbolic surfaces, governed by the same intensity function: $\lambda(\ell) = \frac{\cosh(\ell)-1}{\ell}$.

However, a critical distinction exists. In hyperbolic geometry, the "Collar Lemma" prevents an abundance of short, self-intersecting curves. Metric graphs lack this constraint. Consequently, the authors must distinguish between cycles (paths visiting each edge at most once) and general closed curves.

Figure 4: Comparison between a simple cycle (left) and a more complex walk/curve (right) on a metric ribbon graph.

As proven in Theorem D, without restricting the count to cycles, the expected number of short curves would diverge to infinity, breaking the Poisson convergence. Metric graphs can have an infinite number of short, highly complex, self-intersecting curves hidden in the "thin parts" of the moduli space.

Main Results and Empirical Validation

The core result (Theorems A & B) proves that as $g \to \infty$, the length spectrum of short cycles converges to a Poisson point process. A direct consequence (Corollary C) is that the girth—the length of the absolute shortest cycle—follows a specific non-homogeneous exponential distribution.

To validate these theoretical findings, the authors performed numerical simulations, generating 1,000 random unicellular maps for various genera using Rémy’s algorithm and a Dirichlet distribution for edge lengths.

Figure 6: Empirical distribution of cycle lengths for genus g = 64 compared to the theoretical Poisson intensity curve.

As shown in Figure 6, the simulated data for $g=64$ aligns closely with the predicted theoretical curve, confirming the asymptotic behavior even at finite (though large) genus.

Conclusion and Future Directions

This work establishes a robust link between the combinatorial world of ribbon graphs and the geometric world of hyperbolic surfaces in the high-complexity limit. Future research may explore the Laplacian spectrum (vibrational frequencies) of these graphs or formally define their "local limit" (the Benjamini-Schramm limit), which is conjectured to be an infinite trivalent tree with exponentially distributed edge lengths.

Summary updated

Abstract

Abstract We study the length of short cycles on uniformly random metric maps (also known as ribbon graphs) of large genus using a Teichmüller theory approach. We establish that, as the genus tends to infinity, the length spectrum converges to a Poisson point process with an explicit intensity. This result extends the work of Janson and Louf to the multi-faced case.
BibTeX
@inproceedings{barazer2025length,
  author = {Simon Barazer and Alessandro Giacchetto and Mingkun Liu},
  title = {{Length spectrum of large genus random metric maps}},
  booktitle = {Forum},
  year = {2025},
  doi = {10.1017/fms.2025.31},
}
Figure