Skip to main content
QUICK REVIEW

[Paper Review] Large genus behavior of topological recursion

Bertrand Eynard|arXiv (Cornell University)|May 27, 2019
Mathematical Dynamics and Fractals8 references4 citations
TL;DR

This paper establishes factorial upper bounds on the large-genus growth of topological recursion invariants $ F_g $ for generic regular spectral curves, showing $ |F_g| = O((5g)! r^{-g}) $ under smoothness assumptions. The result implies the generating series $ extstyle rac{1}{eta} ext{Borel transform} $ converges in a disk, setting the stage for resurgence analysis in quantum invariants and enumerative geometry.

ABSTRACT

We show that for a rather generic set of regular spectral curves, the Topological-Recursion invariants F_g grow at most like $O((βg)! r^{-g}) $ with some $r>0$ and $β\leq 5$.

Motivation & Objective

  • To determine the asymptotic large-genus behavior of topological recursion invariants $ F_g $ associated with generic spectral curves.
  • To establish rigorous upper bounds on the growth rate of $ F_g $ as $ g \to \infty $, particularly in terms of factorial decay.
  • To provide conditions under which the generating series $ \sum_g \hbar^{2g-2} F_g $ is asymptotic with factorially bounded coefficients.
  • To lay the foundation for future analysis of resurgence and Borel resummation of the partition function in topological recursion.
  • To quantify the divergence of $ F_g $ as ramification points coalesce, using the radius $ R $ of disjoint spectral discs.

Proposed method

  • Imposes generic smoothness conditions on the spectral curve $ \mathcal{S} = (\Sigma, x, y, B) $, including simple ramification and analyticity of $ y $ near branch points.
  • Introduces local coordinates $ \rho(p) = \sqrt{\prod_{a \in \mathfrak{R}} (x(p) - a)} $ to define a domain $ \Sigma_R $ of disjoint discs around ramification points.
  • Defines key constants $ C $ and $ B $ from the topological recursion kernel $ K $ and bidifferential $ B $, ensuring finiteness under assumptions.
  • Derives recursive bounds on the multidifferentials $ \omega_{g,n} $ via a recursive sequence $ C_{g,n} $, using combinatorial decomposition over stable pairs $ (g_i, n_i) $.
  • Applies a refined estimate involving $ (d+k)^{d+k}/(k^k d^d) $ to control growth in the recursion, leading to factorial-type bounds.
  • Uses contour integration and the $ \omega_{g,1} $-formula to bound $ F_g $, incorporating the $ \Phi $-form and $ \rho $-scaling to derive $ O((5g)! r^{-g}) $ decay.

Experimental results

Research questions

  • RQ1How fast can the topological recursion invariants $ F_g $ grow as $ g \to \infty $ for generic spectral curves?
  • RQ2Under what conditions is the generating series $ \sum_g \hbar^{2g-2} F_g $ asymptotic with factorially bounded coefficients?
  • RQ3Can the Borel transform of the $ F_g $-series be shown to converge in a disk, enabling potential resummation?
  • RQ4What is the precise dependence of $ F_g $ on the distance between ramification points, as measured by $ R $?
  • RQ5Is the growth of $ F_g $ compatible with resurgence, and what are the implications for the full partition function?

Key findings

  • For generic regular spectral curves with simple ramification and analytic $ y $, the invariants satisfy $ |F_g| \leq \tilde{C} C^{2g-2} B^{g-1} \frac{1}{R^{6g-6}} \frac{C_{g,1}}{2g-2} $, with $ \tilde{C} $ depending on $ \Phi $-behavior.
  • A refined bound yields $ |F_g| \leq \tilde{C} \frac{9}{80e} C^{2g-2} B^{g-1} \frac{1}{R^{6g-6}} r^{-g} \frac{(5g-2)!}{2g-2} $, showing at most $ O((5g)!) $ growth.
  • The series $ \sum_g \hbar^{2g-2} F_g $ is asymptotic with factorially bounded coefficients, implying its Borel transform converges in a disk.
  • The bound is sharp in the sense that $ \beta \leq 5 $ is optimal under the assumptions, and $ \beta = 2 $ is typical in known examples.
  • The divergence of $ F_g $ as $ R \to 0 $ (ramification points coalescing) is captured by the $ R^{-6g+6} $ term, confirming known physical intuition.
  • The paper sets the stage for future work on resurgence, suggesting that $ \hat{F}(s) = \sum_g \frac{s^{\beta g}}{(\beta g)!} F_g $ may be analytically continued, leading to instanton corrections of the form $ e^{-\hbar^{-1}} $.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.