Abstract
We introduce and study a new random surface, which we call the hyperbolic Brownian plane and which is the near-critical scaling limit of the hyperbolic triangulations constructed by Curien (Probab Theory Relat Fields 165(3):509–540, 2016). The law of the hyperbolic Brownian plane is obtained after biasing the law of the Brownian plane of Curien and Le Gall (J Theoret Probab 27(4):1249–1291, 2014) by an explicit martingale depending on its perimeter and volume processes studied by Curien and Le Gall (Probab Theory Relat Fields 166(1):187–231, 2016). Although the hyperbolic Brownian plane has the same local properties as those of the Brownian plane, its large scale structure is much different since we prove e.g. that is has exponential volume growth.
This is a preview of subscription content, access via your institution.
Notes
We can find in the ball of radius r of \( \mathbf {T}_{\kappa }\) a number of points at distance at least \(\frac{r}{10}\) from each other that goes to \(+\infty \) as \(r \rightarrow +\infty \), so the sequence \(\big ( \frac{1}{r} B_r(\mathbf {T}_{\kappa }) \big )_{r \ge 1}\) is not tight for the Gromov–Hausdorff topology
References
Abraham, C.: Rescaled bipartite planar maps converge to the Brownian map. Ann. Inst. H. Poincaré Probab. Stat. 52(2), 575–595 (2016)
Abraham, R., Delmas, J.-F., Hoscheit, P.: A note on the Gromov–Hausdorff–Prokhorov distance between (locally) compact metric measure spaces. Electron. J. Probab. 18(14), 1–21 (2013)
Addario-Berry, L., Albenque, M.: The scaling limit of random simple triangulations and random simple quadrangulations. arXiv:1306.5227
Addario-Berry, L., Broutin, N., Goldschmidt, C.: The continuum limit of critical random graphs. Probab. Theory Relat. Fields 152(3–4), 367–406 (2012)
Addario-Berry, L., Broutin, N., Goldschmidt, C., Miermont, G.: The scaling limit of the minimum spanning tree of the complete graph. arxiv:1301.1664 (2013)
Aldous, D., Lyons, R.: Processes on unimodular random networks. Electron. J. Probab. 12(54), 1454–1508 (2007). (electronic)
Angel, O.: Growth and percolation on the uniform infinite planar triangulation. Geom. Funct. Anal. 13(5), 935–974 (2003)
Angel, O., Ray, G.: Classification of half-planar maps. Ann. Probab. 43(3), 1315–1349 (2015)
Angel, O., Schramm, O.: Uniform infinite planar triangulations. Comm. Math. Phys. 241(2–3), 191–213 (2003)
Beltran, J., Le Gall, J.-F.: Quadrangulations with no pending vertices. Bernoulli 19, 1150–1175 (2013)
Benjamini, I., Curien, N.: Ergodic theory on stationary random graphs. Electron. J. Probab. 17(93), 1–20 (2012)
Bertoin, J.: Lévy processes. Cambridge Tracts in Mathematics, vol. 121. Cambridge University Press, Cambridge (1996)
Bettinelli, J., Jacob, E., Miermont, G.: The scaling limit of uniform random plane maps, via the Ambjørn-Budd bijection. Electron. J. Probab. 19 (2014)
Budd, T., Curien, N.: Growth in infinite planar maps with high degrees. arxiv:1602.01328 (2016)
Burago, D., Burago, Y., Ivanov, S.: A Course in Metric Geometry. Graduate Studies in Mathematics, vol. 33. American Mathematical Society, Providence (2001)
Curien, N.: Planar stochastic hyperbolic triangulations. Probab. Theory Relat. Fields 165(3), 509–540 (2016)
Curien, N., Le Gall, J.-F.: The Brownian plane. J. Theoret. Probab. 27(4), 1249–1291 (2014)
Curien, N., Le Gall, J.-F.: First-passage percolation and local perturbations on random planar maps. arXiv:1511.04264, (2015)
Curien, N., Le Gall, J.-F.: The hull process of the Brownian plane. Probab. Theory Relat. Fields 166(1), 187–231 (2016)
Curien, N., Le Gall, J.-F.: Scaling limits for the peeling process on random maps. Ann. Inst. H. Poincaré Probab. Stat. 53(1), 322–357 (2017)
den Hollander, F.: Probability Theory: The Coupling Method. http://websites.math.leidenuniv.nl/probability/lecturenotes/CouplingLectures.pdf
Duminil-Copin, H., Garban, C., Pete, G.: The near-critical planar FK-Ising model. Comm. Math. Phys. 326(1), 1–35 (2014)
Garban, C., Pete, G., Schramm, O.: The Scaling Limits of Near-critical and Dynamical Percolation. arXiv:1305.5526, (2013)
Kesten, H.: Scaling relations for \(2\)D-percolation. Comm. Math. Phys. 109(1), 109–156 (1987)
Krikun, M.: Explicit enumeration of triangulations with multiple boundaries. Electron. J. Combin., 14(1):Research Paper 61 (electronic) (2007)
Le Gall, J.-F.: Random geometry on the sphere. In: Proceedings of the ICM (2014)
Le Gall, J.-F.: Uniqueness and universality of the Brownian map. Ann. Probab. 41, 2880–2960 (2013)
Le Gall, J.-F.: The Brownian cactus II: upcrossings and local times of super-Brownian motion. Probab. Theory Relat. Fields 162(1), 199–231 (2014)
Miermont, G.: Aspects of random maps. preprint, http://perso.ens-lyon.fr/gregory.miermont/coursSaint-Flour.pdf
Miermont, G.: The Brownian map is the scaling limit of uniform random plane quadrangulations. Acta Math. 210(2), 319–401 (2013)
Miller, J., Sheffield, S.: An axiomatic characterization of the Brownian map. arXiv:1506.03806, (2015)
Nolin, P., Werner, W.: Asymmetry of near-critical percolation interfaces. J. Am. Math. Soc. 22(3), 797–819 (2009)
Stephenson, R.: Local convergence of large critical multi-type Galton–Watson trees and applications to random maps. J. Theoret. Probab., 1–47, (2016)
Wen, Y.: The Brownian plane with minimal neck baby universe. arXiv:1511.01028 (2015)
Acknowledgements
I thank Nicolas Curien for suggesting me to study this object, and for carefully reading many earlier versions of this manuscript. I also thank the anonymous referee for his useful comments. I acknowledge the support of ANR Liouville (ANR-15-CE40-0013) and ANR GRAAL (ANR-14-CE25-0014).
Author information
Authors and Affiliations
Corresponding author
Appendix A: Proof of Proposition 10
Appendix A: Proof of Proposition 10
The goal of this appendix is to prove Proposition 10. We note that similar ideas to those below appear in Sect. 2 of [31], and more precisely in the proofs of Proposition 2.17 and 2.18. In particular, Lemma 30 is essentially proved in the proof of Proposition 2.18 there. If A is a subset of a metric space X and \(\varepsilon >0\), we will write \(A^{\varepsilon }\) for the union of all the open balls of radius \(\varepsilon \) centered at an element of A. Note that this is an open subset of X. We also recall that \(B_r(X)\) and \(\overline{B_r}(X)\) are respectively the closed ball and the hull centered at the root of X. In particular, they are both closed subsets of X. To prove Proposition 10, we will also need several times the following definition.
Definition 27
Let (X, d) be a metric space, \(x,y \in X\) and \(\varepsilon >0\). An \(\varepsilon \)-chain from x to y is a finite sequence of points \((z_i)_{0 \le i \le k}\) of X such that \(z_0=x\), \(z_k=y\) and \(d(z_i,z_{i+1}) \le \varepsilon \) for all \(0 \le i \le k-1\).
Lemma 28
Under the assumptions of Proposition 10, the application \(s \rightarrow \overline{B_s}(X)\) is continuous at r for the Hausdorff distance on the set of the compact subsets of \(\overline{B_{r+1}}(X)\).
Proof
Let \(\delta >0\). It is enough to show that, for some \(\varepsilon >0\), we have:
-
(a)
\(\overline{B_{r+\varepsilon }}(X) \subset \overline{B_r}(X)^{\delta }\),
-
(b)
\(\overline{B_r}(X) \subset \overline{B_{r-\varepsilon }}(X)^{\delta }\).
We start with the first point. Let \(A=\Big ( \bigcap _{\varepsilon >0} \overline{B_{r+\varepsilon }}(X) \Big ) \backslash \overline{B_r}(X)\). If \(x \in A\), then there is a geodesic from x to a point \(y \in X \backslash \overline{B_{r+1}}(X)\) that stays outside \(B_r(X)\). However, \(\gamma \) has to intersect \(B_{r+\varepsilon }(X)\) for all \(\varepsilon >0\). This is clearly impossible, so \(A=\emptyset \). In particular, the decreasing intersection of the compact sets \(\overline{B_{r+1/n}}(X) \backslash \overline{B_r}(X)^{\delta }\) is empty, so one of these compact sets is empty, which proves item (a).
For item b), let \(A'=\big ( \overline{B_r}(X)\backslash B_r(X) \big ) \backslash \Big ( \bigcup _{\varepsilon >0} \overline{B_{r-\varepsilon }}(X) \Big )\). By assumption (iii) we have \(\mu (A')=\emptyset \) and \(A'\) is open, so by assumption (ii) we get \(A'=\emptyset \). This implies \( \overline{B_r}(X)\backslash B_r(X) \subset \bigcup _{n \ge 0} \overline{B_{r-1/n}}(X)^{\delta }\). Moreover, we have \(B_r(X) \subset \overline{B_{r-1/n}}(X)^{\delta }\) for \(\frac{1}{n} < \delta \). Hence, the increasing family of open sets \(\overline{B_{r-1/n}}(X)^{\delta }\) covers the compact space \(\overline{B_r}(X)\), so there is an \(\varepsilon >0\) such that \(\overline{B_r}(X) \subset \overline{B_{r-\varepsilon }}(X)^{\delta }\).
Remark 29
Note that the first point is true even without assumptions (ii) and (iii).
Lemma 30
Let \(\varepsilon >0\). Let \((X, \rho , a)\) and \((Y, \rho , b)\) be two bipointed connected, compact subsets of a locally compact metric space (Z, d). Assume that the Hausdorff distance between X and Y is less than \(\varepsilon \) and that \(d(a,b) \le \varepsilon \). Then, for all r such that \(r+4 \varepsilon < d(\rho ,b)\), we have
-
1)
\(\overline{B_r}(X)^{\varepsilon } \cap Y \subset \overline{B_{r+4\varepsilon }}(Y)\),
-
2)
\(\overline{B_r}(X) \subset \overline{B_{r+4\varepsilon }}(Y)^{\varepsilon }\).
Illustration of the proof of Lemma 30
Proof
We first notice that the connectedness of X implies that, for any two points x and \(x'\) in X, there is an \(\varepsilon \)-chain in X from x to \(x'\). The same is true for Y.
Let y be a point in \(\overline{B_r}(X)^{\varepsilon } \cap Y\). We want to show that \(y \in \overline{B_{r+4\varepsilon }}(Y)\). We can assume \(d(\rho , y)>r+4\varepsilon \) (if it is not the case, then \(y \in \overline{B_{r+4\varepsilon }}(Y)\) is obvious). Let \((z_i)_{0 \le i \le k}\) be an \(\varepsilon \)-chain from y to b in Y. We know that \(d(a,b) \le \varepsilon \) and that there is a point \(x \in X\) such that \(d(x,y)\le \varepsilon \). We write \(w_0=x\), \(w_k=a\) and, for all \(1 \le i \le k-1\), we take \(w_i \in X\) such that \(d(w_i,z_i) \le \varepsilon \) (see Fig. 1 for an illustration). For all i, we have \(d(w_i,w_{i+1}) \le d(w_i,z_i)+d(z_i,z_{i+1})+d(z_{i+1},w_{i+1}) \le 3 \varepsilon \), so \((w_i)\) is a \(3\varepsilon \)-chain from x to a in X. Since \(x \in \overline{B_r}(X)\), there must be at least one i such that \(d(\rho ,w_i) \le r+3\varepsilon \), which implies \(d(\rho ,z_i) \le r+4\varepsilon \). Hence, every \(\varepsilon \)-chain from y to b has at least one point at distance from the root less than \(r+4\varepsilon \), so there is no \(\varepsilon \)-chain from y to b in \(Y \backslash B_{r+4 \varepsilon }(Y)\). This implies that y and b do not lie in the same connected component of \(Y \backslash B_{r+4 \varepsilon }(Y)\), so \(y \in \overline{B_{r+4\varepsilon }}(Y)\), which proves the first point of the lemma.
The second point is easily obtained from the first one: if \(x \in \overline{B_r}(X)\), then there is a \(y \in Y\) such that \(d(x,y) \le \varepsilon \). By the first point, we have \(y \in \overline{B_{r+4\varepsilon }}(Y)\), so \(x \in \overline{B_{r+4\varepsilon }}(Y)^{\varepsilon }\). \(\square \)
Proof of Proposition 10
First, we need to prove that the radii of the \(\overline{B_r}(X_n)\) are bounded. Let \(R=\max \{ d(\rho ,x) | x \in \overline{B_{r+4}}(X)\}\) be the radius of \(\overline{B_{r+4}}(X)\). For n large enough, we have
We write \(X'=B_{R+3}(X)\) and \(X'_n=B_{R+3}(X_n)\). The above inequality means that we can embed \(X'\) and \(X'_n\) isometrically in the same space (Z, d) in such a way that \(X'_n \subset (X')^1\) and \(\rho _n=\rho \).
Let \(b \in X'_n \backslash B_{R+2}(X_n)\) and \(a \in X'\) such that \(d(a,b) \le 1\). We have \(d(a,\rho ) \ge R+1\), so \(a \notin \overline{B_r}(X)\) and \(\overline{B_r}(X)=\overline{B_r}(X')\) is the hull of radius r with center \(\rho \) with respect to a in \(X'\). By item 2) of Lemma 30 for \(\varepsilon =1\), we have \(\overline{B_r}(X'_n) \subset \overline{B_{r+4}}(X')^1\). In particular, the radius of \(\overline{B_r}(X'_n)\) is less than \(R+1\). This means that only one connected component of \(X' \backslash B_r(X'_n)\) contains points at distance greater than \(R+1\) from the root. In other words, the radius of \(\overline{B_r}(X_n)\) is at most \(R+1\).
We now move on to the proof of our proposition. Let \(\delta >0\). By Lemma 28, there is \(\varepsilon >0\) such that \(\overline{B_{r+4 \varepsilon }}(X) \subset \overline{B_r}(X)^{\delta }\) and \(\overline{B_r}(X) \subset \overline{B_{r-4 \varepsilon }}(X)^{\delta }\). For n large enough, we have \(d_{GHP}(X',X'_n) \le \varepsilon \). This means that we can embed \(X'\) and \(X'_n\) in the same space Z in such a way that
-
(a)
\(\rho =\rho _n\),
-
(b)
\(X'_n \subset (X')^{\varepsilon }\),
-
(c)
\(X' \subset (X'_n)^{\varepsilon }\),
-
(d)
for every \(A \subset X'_n\) that is measurable we have \(\mu _n(A) \le \mu (A^{\varepsilon })+\varepsilon \),
-
(e)
for every \(B \subset X'\) that is measurable we have \(\mu (B) \le \mu _n(B^{\varepsilon })+\varepsilon \).
This embedding provides a natural way to embed the measured metric spaces \(\overline{B_r}(X_n)\) and \(\overline{B_r}(X)\) in Z. We will deduce an upper bound for the GHP distance between these two hulls.
For all \(y \in \overline{B_r}(X_n)\) there is an \(x \in X'\) such that \(d(x,y) \le \varepsilon \). By Lemma 30 we have \(x \in \overline{B_{r+4\varepsilon }}(X)\), so by Lemma 28 and our choice of \(\varepsilon \), we have \(x \in \overline{B_r}(X)^{\delta }\). This proves \(\overline{B_r}(X_n) \subset \overline{B_r}(X)^{\delta +\varepsilon }\).
Similarly, let \(x \in \overline{B_r}(X)\). We have \(x \in \overline{B_{r-4\varepsilon }}(X)^{\delta }\) by Lemma 28 and our choice of \(\varepsilon \). Let \(z \in \overline{B_{r-4\varepsilon }}(X)\) be such that \(d(x,z) \le \delta \). There is a \(y \in X'_n\) such that \(d(y,z) \le \varepsilon \). By Lemma 30 we have \(y \in \overline{B_{r-4\varepsilon +4\varepsilon }}(X_n)\) and \(d(x,y) \le \delta +\varepsilon \), which proves \(\overline{B_r}(X) \subset \overline{B_r}(X_n)^{\delta +\varepsilon }\). Hence, in our embedding, the Hausdorff distance between \(\overline{B_r}(X)\) and \(\overline{B_r}(X_n)\) is less than \(\varepsilon +\delta \).
For all \(A \subset \overline{B_r}(X_n)\) measurable, we have \(\mu _n(A) \le \mu (A^{\varepsilon }) +\varepsilon =\mu \big ( A^{\varepsilon } \cap X' \big ) +\varepsilon \). By Lemma 30 we have the inclusion \(A^{\varepsilon } \cap X' \subset A^{\varepsilon } \cap \overline{B_{r+4\varepsilon }}(X)\), so we get
where we recall that \(V(s)=\mu (\overline{B_s}(X))\) for all s.
Similarly, for all \(B \subset \overline{B_r}(X)\) measurable, we have
where the last inequality uses Lemma 30.
Hence, our embedding of \(\overline{B_r}(X)\) and \(\overline{B_r}(X_n)\) gives the following bound for n large enough:
By assumption (iii) in the statement of the proposition, the right-hand side can be made arbitrarily small, which proves the first point of Proposition 10. The second point is an obvious consequence of the first one.
Finally, in the case of compact, bipointed metric spaces \(\big ( (X_n, d_n), x_n,y_n, \mu _n \big )\) and \(\big ( (X, d), x,y, \mu \big )\) with \(r <d(x,y)\), the above proof still works. The first part of the proof (i.e. proving that the radii of the \(\big ( \overline{B_r}(X_n) \big )_{n \ge 0}\) are bounded) is not necessary anymore. We may apply the second part of the proof directly to X and \(X_n\) instead of the compact subsets \(X'\) and \(X'_n\). Note that if \(r>d(x,y)\), then \(\overline{B_r}(X)=X\) and \(\overline{B_r}(X_n)=X_n\) for n large enough, so the conclusion is immediate. \(\square \)
Rights and permissions
About this article
Cite this article
Budzinski, T. The hyperbolic Brownian plane. Probab. Theory Relat. Fields 171, 503–541 (2018). https://doi.org/10.1007/s00440-017-0785-x
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00440-017-0785-x
Mathematics Subject Classification
- 60F17
- 60D05
- 60C05
