Scaling limit of first passage percolation geodesics on planar maps
Emmanuel Kammerer · 2024
Read Scaling limit of first passage percolation geodesics on planar maps by Emmanuel Kammerer (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-02666.txt.
Introduction
Emmanuel Kammerer, a solo author, presents a 2024 math paper on random planar maps. First passage percolation puts random travel times on a map's edges. Geodesics are the fastest routes under those times. Kammerer finds the scaling limit of geodesics to the root as maps grow huge. The count of faces along a geodesic has its own limit law. That law lets two distance notions be compared: percolation distance and dual graph distance.
An upper bound for the diameter of large random maps follows. The whole tree of geodesics converges to a flow of jumping diffusions that merge as they run. From that flow, candidate limit spaces for high-degree maps are built. Probabilists get a rare full picture of route geometry on random surfaces. The peeling exploration, run backward in time, is the key tool on display.
Catalog record
| CREATOR | Emmanuel Kammerer |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 174 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-02666
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Scaling limit of first passage percolation geodesics on planar maps Emmanuel Kammerer Abstract We establish the scaling limit of the geodesics to the root for the first passage percolation distance on random planar maps. We first describe the scaling limit of the number of faces along the geodesics. This result enables us to compare the metric balls for the first passage percolation and the dual graph distance. It also enables us to give an upper bound for the diameter of large random maps. Then, we describe the scaling limit of the tree of first passage percolation geodesics to the root via a stochastic coalescing flow of pure jump diffusions. Using this stochastic flow, we also construct some random metric spaces which we conjecture to be the scaling limits of random planar maps with high degrees. The main tool in this work is a time-reversal of the uniform peeling exploration. Figure 1: Simulation of four trajectories of the stochastic coalescing flow of diffusions with jumps which arises as the scaling limit of first-passage percolation geodesics. 1 Introduction 1.1 Random Boltzmann maps We study the first-passage percolation distance on random Boltzmann planar maps of type a\in(3/2,5/2] which correspond to the dual of the stable maps of [ LGM11 ] when a\in(3/2,5/2) and to random planar maps in the universality class of the Brownian sphere when a=5/2 . We give applications to the (dual) graph distance and establish the scaling limit of geodesics to the root. A planar map is a connected planar graph which is embedded in the sphere and seen up to orientation-preserving transformations. Our planar maps are equipped with a distinguished oriented edge which is called the root edge. The face f_{r} which is on the right of the root edge is called the root face. We restrict our attention to bipartite planar maps, i.e. such that each face has an even degree. We denote by \mathcal{M} the set of finite bipartite planar maps and for all \ell\geq 1 we denote by \mathcal{M}^{(\ell)} the set of bipartite planar maps whose root face has degree 2\ell . See Figure 2 for an example. A general way to pick a map at random consists in assigning positive weights to each face according to its degree. Following [ MM07 ] , let {\bf q}=(q_{k})_{k\geq 1} be a non-zero sequence of non-negative real numbers. The weight of a map \mathfrak{m}\in\mathcal{M} is defined by w_{\bf q}(\mathfrak{m})=\prod_{f\in\mathrm{Faces}(\mathfrak{m})\setminus\{f_{r}\}}q_{\deg(f)/2}. We then define, for all \ell\geq 1 , the partition function W^{(\ell)}=\sum_{\mathfrak{m}\in\mathcal{M}^{(\ell)}}w_{\bf q}(\mathfrak{m}). When the partition function is finite, the sequence \bf q is called admissible. When {\bf q} is admissible, we define the Boltzmann probability measure \mathbb{P}^{(\ell)} on \mathcal{M}^{(\ell)} by setting \forall\mathfrak{m}\in\mathcal{M}^{(\ell)},\qquad\mathbb{P}^{(\ell)}(\{\mathfrak{m}\})=\frac{w_{\bf q}(\mathfrak{m})}{W^{(\ell)}}. A random map of law \mathbb{P}^{(\ell)} is called a random Boltzmann map of perimeter 2\ell . The geometry of random Boltzmann maps heavily depends on the asymptotic behaviour of the partition function W^{(\ell)} as \ell\to\infty . It is known that there exists a constant c_{\bf q}>0 such that W^{(\ell)}=O(c_{\bf q}^{\ell}\ell^{-3/2})\qquad\text{and}\qquad W^{(\ell)}=\Omega(c_{\bf q}^{\ell}\ell^{-5/2}) as \ell\to\infty . See Section 5.3 of [ Cur23 ] . As in the lecture notes [ Cur23 ] , we say that the weight sequence \bf q is subcritical (or of type a=3/2 ) if there exists a constant p_{\bf q}>0 such that W^{(\ell)}\mathop{\sim}\limits_{\ell\to\infty}\frac{p_{\bf q}}{2}c_{\bf q}^{\ell+1}\ell^{-3/2}, while the sequence \bf q is said to be critical generic (or of type a=5/2 ) if there exists p_{\bf q}>0 such that W^{(\ell)}\mathop{\sim}\limits_{\ell\to\infty}\frac{p_{\bf q}}{2}c_{\bf q}^{\ell+1}\ell^{-5/2}. When q_{k}=0 for k large enough, i.e. when the face degrees are bounded, and \bf q is admissible, then either \bf q is subcritical, or it is generic critical. To obtain intermediate exponents, one has to favor the appearance of high degrees by choosing \bf q so that the weights decay “slowly enough”. We say that \bf q is critical non-generic of type a with 3/2<a<5/2 if there exists p_{\bf q}>0 such that W^{(\ell)}\mathop{\sim}\limits_{\ell\to\infty}\frac{p_{\bf q}}{2}c_{\bf q}^{\ell+1}\ell^{-a}. (1.1) A random Boltzmann map associated with a critical non-generic weight sequence \bf q of type a will also be called a stable map with parameter a-1/2 . Examples of such sequences \bf q are exhibited in Lemma 6.1 of [ BC17 ] . The study of random Boltzmann maps with \bf q being critical non-generic is motivated by strong connections with random maps coupled with models of statistical mechanics. More precisely, if one considers a random map coupled with a model of statistical mechanics, then, at criticality, the cluster of the root face, which is also called the gasket, is a random Boltzmann map with a critical non-generic weight sequence \bf q . See e.g. [ BCM19 ] for Bernoulli percolation, [ AM22 ] for the Ising model and [ BBG12 ] for the O(n) loop model. Random planar maps coupled with models of statistical mechanics were first studied by physicists as models of statistical mechanics in quantum gravity. See e.g. [ EK95 ] and the references therein for the O(n) loop model. Scaling limits of random maps coupled with models of statistical mechanics are conjectured to be described by Liouville quantum gravity, first introduced in physics in [ Pol81 ] , and by an independent conformal loop ensemble, defined in [ She09 , SW12 ] . See e.g. Conjecture 2.1 in [ HL24 ] for a precise conjecture in the case the O(n) loop model and see [ GHS21 ] for a result solving the analogous problem in the case of Bernoulli percolation. In what follows, we assume that ( 1.1 ) holds for some a\in(3/2,5/2] , i.e. that \bf q is either critical generic or critical non-generic. If one conditions a random map of law \mathbb{P}^{(\ell)} to have n vertices and lets n\to\infty , then the law of the ball of fixed radius R centered at the root face f_{r} converges for all R>0 . This enables us to define an infinite random map called the local limit, whose law is given by Theorem 7.1 of [ Cur23 ] and is denoted by \mathbb{P}^{(\ell)}_{\infty} . A random map of law \mathbb{P}^{(\ell)}_{\infty} is called an infinite random Boltzmann map of perimeter 2\ell . When \mathfrak{m} is a (possibly infinite) map, we denote by \mathfrak{m}^{\dagger} the dual map associated with \mathfrak{m} which is obtained by switching the roles of the faces and vertices and saying that two vertices of \mathfrak{m}^{\dagger} are connected by an edge if the corresponding faces of \mathfrak{m} are adjacent. Different distances can then be studied, notably: • The (primal) graph distance d_{\mathrm{gr}} on the vertices of \mathfrak{m} , obtained by assigning a length 1 to each edge; • The dual graph distance d_{\mathrm{gr}}^{\dagger} on the faces of \mathfrak{m} , which are the vertices of \mathfrak{m}^{\dagger} , obtained by assigning a length 1 to each dual edge. This distance corresponds to the graph distance on the dual map \mathfrak{m}^{\dagger} ; • The first-passage percolation (fpp) distance d^{\dagger}_{\mathrm{fpp}} on the faces of \mathfrak{m} , obtained by assigning i.i.d. exponential random edge lengths of parameter 1 to the dual edges. The primal graph distance is now well understood in the generic case a=5/2 since the works of Le Gall [ LG13 ] and Miermont [ Mie13 ] . The scaling limit of random maps of law \mathbb{P}^{(\ell)} in the case a=5/2 as