The hyperbolic Brownian plane (Q1647936)

The paper under review introduces and to studies the hyperbolic Brownian plane, which is the near-critical scaling limit of the hyperbolic triangulation constructed by \textit{N. Curien} [Probab. Theory Relat. Fields 165, No. 3--4, 509--540 (2016; Zbl 1342.05137)]. The hyperbolic Brownian plane has the same local properties as those of the Brownian plane, and is different from the Brownian plane in its large scale structure. In [Ann. Probab. 43, No. 3, 1315--1349 (2015; Zbl 1354.60010)], \textit{O. Angel} and \textit{G. Ray} characterized all the triangulations of the half-plane enjoying a spatial Markov property and discovered a new family of triangulations of the half-plane with hyperbolic flavor. The triangulations \(T_s\) for \(k < k_{\varepsilon}\) are called type-H planar stochastic hyperbolic triangulation (PSHT) if there is a one-parameter family \(T_k, 0<k<k_{\varepsilon}\), of Markovian random triangulations of the plane. At the critical value \(k=k_{\varepsilon}\) the random triangulation \(T_{k_{\varepsilon}}\) is the uniform infinite planar triangulation (UIPT) of \textit{O. Angel} and \textit{O. Schramm} [Commun. Math. Phys. 241, No. 2--3, 191--213 (2003; Zbl 1098.60010)]. The main tool for the scale limit at the right speed as the distance renormalized consists of the absolute continuity between the hyperbolic triangulations and UIPT. The main theorem (Theorem 1) indicates with the parameter \(\lambda_n = \lambda_{\varepsilon}\), the convergence \(\frac{1}{T} T_{\lambda_n} \to {\mathcal P}^h\) to a random locally compact metric space (as hyperbolic Brownian plane), and the distribution is characterized by its characteristic function. Theorem 2 extends the convergence toward the Brownian plane in the sense of local Gromov Hausdorff Prokhornov, and Theorem 3 defines the perimeter and the volume of the Brownian Hull of radius \(r\). Section 1 prepares enumeration and type-I PSHT, and type-I triangulation of a \(p\)-gon, which is a planar map equipped with a distinguished oriented edge. A combinatorial estimate of the set of type-I triangulations of \(p\)-gon with \(n\) inner vertices is given in Lemma 2 using the Stirling formula and \textit{N. Curien} and \textit{J.-F. Le Gall} [Ann. Inst. Henri Poincaré, Probab. Stat. 53, No. 1, 322--357 (2017; Zbl 1358.05255)]. The author constructs an analog of the hyperbolic triangulations in the case of type I triangulations. Proposition 5 shows that there is a critical \(\lambda_{\varepsilon}\) such that there is no \(\lambda\)-Markovian type I triangulation for \(\lambda> \lambda_{\varepsilon}\) and a unique \(\lambda\)-Markovian type I triangulation in the distribution sense. Section 2 starts with the recall of the Gromov-Hausdorff-Prokhorov (GHP) distance and states that measurable functions of the closed hull converge for the GHP distance to the measurable functions on the closed hull in a GHP convergent metric space if there is a GHP convergence in the local GHP distance topology. The main tools come from theorems of \textit{N. Curien} and \textit{J.-F. Le Gall} [J. Theor. Probab. 27, No. 4, 1249--1291 (2014; Zbl 1305.05208)] for quadrangulations and for the Gromov-Hausdorff distance. The joint convergence in Theorem 2 is given in Proposition 13 and deduced from Proposition 11, where Proposition 11 is proved by Proposition 12 and two small modification of Theorem 2 in Curien and Le Gall [2014, loc. cit.]. The associated function \(\phi (p, v)\) and \(\phi_{\lambda}(p, v)\) in subsection 2.4 is used for the convergence issue. For a triangulation \(t\) with a simple hole of perimeter \(p^{'}\) and total vertices \(v^{'}\) with \(\lambda \in (0, \lambda_c]\), let the function \(\phi_{\lambda_n} (p_n^{'}, v_n^{'}) \to \phi (p, v)\) for sequences \(p_n^{'}/n^2 \to \frac{3p}{2}\) and \(v_n^{'}n^4 \to 3v\) and Markovian \(T_{\lambda_n}\) and \(T\). This proposition 16 follows from the definition of the type-I PSHT. Hence, Theorem 1 indicates that \(\frac{T_{\lambda_n}}{n} \to {\mathcal P}^h\) as \(n \to \infty\) in the sense of the local Gromov-Hausdorff-Prokhorov distance for a random locally compact metric space (hyperbolic Brownian plane) with the specified distribution. The proof of Theorem 1 follows from the convergence in Theorem 1 and Proposition 16 and the Kolmogorov extension theorem. Section 3 devotes to study properties of the hyperbolic Brownian plane \({\mathcal P}^h\). The hyperbolic Brownian plane \({\mathcal P}^h\) has Hausdorff dimension 4 and is homeomorphic to \(\mathbb R^2\) almost surely, and carries a property that is continuum analog of unimodularity for random graphs in Proposition 21, and further identifies two processes \(P^h\) and \(V^h\) in a more convenient way to express in Theorem 3. The proof of Theorem 3 is given in this section by specifying the pure jump process \(V\) and previous Theorem 3. For a measurable enumeration \(t_i\) of the jumps \(S\), the process \(M_t = e^{-S_t} \prod_{t_i \leq t} (1+ 2 |\Delta S_{t_i}| )e^{ - 2 |\Delta S_{t_i}| }\) is martingale with respect to the natural filtration associated to \(S\). Theorem 3 is proved by constructing the extension time and absolute continuity relation for conditioned version of \(Z\) and \(Z^h\) and showing that they share the same one-dimensional marginals and the Laplace transform of variables that are computed in [\textit{N. Curien} and \textit{J.-F. Le Gall}, Probab. Theory Relat. Fields 166, No. 1--2, 187--231 (2016; Zbl 1347.05226)]. Asymptotics for the perimeter and volume processes are given in the last subsection 3.4. The appendix presents a complete proof of a technique Proposition 10 which are in line similar to the proofs of Proposition 2.17 and 2.18.