On the minimal diameter of hyperbolic surfaces Thomas Budzinski (UBC) Joint work with Nicolas Curien (Université Paris–Saclay) and Bram Petri (Sorbonne Université) October 23rd 2019 UBC Probability Seminar Thomas Budzinski Minimal diameter of hyperbolic surfaces
Introduction Goal: study the minimal possible diameter of hyperbolic surfaces with high genus g : asymptotic to 1 × log g . Small diameter ≈ highly connected objects. Random (say, 3-regular) graphs are also very connected: probabilistic method , very common in combinatorics. Probabilistic method in hyperbolic geometry. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Program I The diameter of 3-regular random graps II Notions of hyperbolic geometry III The diameter of hyperbolic surfaces IV Ideas of the proof V Perspectives Thomas Budzinski Minimal diameter of hyperbolic surfaces
The diameter of 3-regular random graphs G n obtained from n vertices with 3 half-edges each, by matching the half-edges uniformly at random (connected with proba 1 − O ( 1 / n ) ). Diameter: maximal graph distance between two vertices. Theorem (Bollobas–Fernandez de la Vega, 1982) diam ( G n ) ( P ) − n → + ∞ 1 . − − − → log 2 n Lower bound: a ball of radius r has size at most 3 × 2 r , so the diameter is ≥ log 2 n . Thomas Budzinski Minimal diameter of hyperbolic surfaces
Diameter of random graphs: proof Upper bound: it is enough to prove that for any two fixed vertices v 1 , v 2 : � 1 � P (( 1 + ε ) log 2 n ≤ d G n ( v 1 , v 2 ) < + ∞ ) = o . n 2 Explore balls of radius r = 1 + ε 2 log 2 n around v 1 and v 2 , and try to connect them. v 1 ? If no "bad step", we would have | ∂ B r ( v 1 ) | = 3 × 2 r = 3 n 1 + ε 2 . Thomas Budzinski Minimal diameter of hyperbolic surfaces
Diameter of random graphs: proof But P ( bad step at time i ) ≤ i + 2 3 n , with independence over i . � 1 � Consequence: with probability 1 − o : n 2 O ( 1 ) bad steps in the ball of radius 1 − ε 2 log 2 n bad steps, � � 1 + ε bad steps between distances 1 − ε o n 2 log 2 n and 2 1 + ε 2 log 2 n , 1 + ε so | ∂ B r ( v 1 ) | ≥ δ n w.h.p., and the same is true around v 2 . 2 B r ( v 1 ) B r ( v 2 ) Thomas Budzinski Minimal diameter of hyperbolic surfaces
Diameter of random graphs: proof If B r ( v 1 ) ∩ B r ( v 2 ) � = ∅ , we are done. If not, each loose half-edge on ∂ B r ( v 1 ) has probability ≥ δ n − 1 − ε | ∂ B r ( v 2 ) | to be connected to B r ( v 2 ) . So 2 n 1 + ε � n 2 � 1 − δ n − 1 − ε P ( B r ( v 1 ) and B r ( v 2 ) not directly linked ) ≤ 2 ≤ exp ( − δ n ε ) � 1 � = o , n 2 and d G n ( v 1 , v 2 ) ≤ 2 r + 1 ≤ ( 1 + ε ) log 2 n with very high probability. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Hyperbolic geometry The hyperbolic plane H can be seen as the unit disk, equipped with the metric 4 d x 2 d s 2 = 1 − | x | 2 . Curvature : | B ε ( x ) | = πε 2 − π 12 ε 4 K ( x ) + o ( ε 4 ) . Riemann uniformization theorem: H is the unique simply connected surface with constant curvature equal to − 1. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Compact hyperbolic surfaces A compact hyperbolic surface S is a 2 d manifold equipped with a Riemannian metric with constant curvature − 1. We consider closed surfaces, i.e. no boundary. � S K ( x ) d x = 2 π ( 2 − 2 g ) , where g is Gauss–Bonnet formula: the genus of the surface, i.e. the number of holes. So g ≥ 2. S Equivalent definitions: S is locally isometric to H , S is a quotient of H (by a nice enough group action), S is a surface equipped with a conformal structure. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Pairs of pants Existence, but no uniqueness: for g ≥ 2, hyperbolic metrics on a genus g surface form a ( 6 g − 6 ) -dimensional space M g called the moduli space . One way to build a lot of them is to use pants . For any ℓ 1 , ℓ 2 , ℓ 3 ≥ 0, there is a unique surface isomorphic to the sphere minus 3 disjoint disks, such that: the boundaries of the three disks are closed geodesics with lengths ℓ 1 , ℓ 2 , ℓ 3 ; the curvature is − 1 outside of the boundary. ℓ 1 ℓ 2 ℓ 3 Thomas Budzinski Minimal diameter of hyperbolic surfaces
Gluings of pants By gluing 2 g − 2 pairs of pants such that the lengths of the boundaries match two by two, we can build many hyperbolic surfaces. 6 g − 6 degrees of freedom: 3 g − 3 for the lengths of the cycles, and 3 g − 3 for the twists. Conversely, every hyperbolic surface of genus g can be cut by 3 g − 3 closed geodesics into 2 g − 2 pairs of pants. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Interesting quantities for hyperbolic surfaces Given a hyperbolic surface S , several natural quantities to look at, and try to optimize over the moduli space: diameter, spectral gap (eigenvalues of the Laplacian), Cheeger constant (isoperimetric inequalities), systole (length of the smallest closed geodesics). All of these measure the "connectivity" of the surface. In the context of hyperbolic surfaces, non-optimal bounds (constant factors) often obtained via arithmetic constructions [Brooks, Buser, Kim, Sarnak...] . A typical graph is very connected, so random graphs (like uniform 3-regular graphs) are close to optimal for these quantities. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Diameter of hyperbolic surfaces Diameter: maximal distance between two points of S . Easy: sup S ∈ M g diam ( S ) = + ∞ . log 1 ε ε − → × 2 ε ε Lower bound for the minimal diameter: volume growth argument By Gauss-Bonnet, Area ( S ) = 2 π ( 2 g − 2 ) . In H , the area of balls is Area ( B r ( x )) = 2 π ( cosh ( r ) − 1 ) . S is a quotient of H , so Area ( B r ( x )) ≤ 2 π ( cosh ( r ) − 1 ) in S . So if B r ( x ) covers S , then cosh ( r ) − 1 ≥ 2 g − 2, so S ∈ M g diam ( S ) ≥ cosh − 1 ( 2 g − 1 ) = log g + O ( 1 ) . inf Best lower bound [Bavard 1996] : also log g + O ( 1 ) . Thomas Budzinski Minimal diameter of hyperbolic surfaces
Main theorem Theorem (B.–Curien–Petri, 2019) We have S ∈ M g diam ( S ) = ( 1 + o ( 1 )) log g . min Construction: random gluing of pants! Start from 2 g − 2 pants with perimeters ( a , a , a ) , and glue the 6 g − 6 holes uniformly at random to obtain S g , a . Twist 0: the "centers" of two neighbour pants have the same projections on the boundary. We show diam ( S g , a ) ∼ 1 β a log g w.h.p., where β a < 1, but β a → 1 as a → ∞ . Thomas Budzinski Minimal diameter of hyperbolic surfaces
A crude bound For a fixed, the diameter of a pair of pants is a constant. Enough to bound distances between the centers of the pants. Quick bound: diam ( S a , g ) ≤ 2 d a diam ( G 2 g − 2 ) n →∞ 2 d a log 2 n . ∼ After computing d a for a → + ∞ , we get ≈ 1 . 38 log g . a c 3 c 2 c 4 c d a a a c 1 Not optimal: sometimes, there is a much shorter path. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Adapting the explorations Instead of using random graphs as a black box, adapt the proof! Adapt the exploration to the hyperbolic metric, instead of the graph distance. Ideal situation: the neighbourhood of one center looks like an infinite tree of pants T a . We need to understand its growth! Thomas Budzinski Minimal diameter of hyperbolic surfaces
Growth of the infinite tree of pants B r : ball of radius r around a center of the pants tree T a . Let | B r | be the number of pants whose center is in B r . Lemma r → + ∞ C a e β a r , where β a → 1 as a → + ∞ . We have | B r ( T a ) | ∼ Sketch of proof: pants can be decomposed in two right-angled hexagons. a a / 2 b b b b − → a / 2 a / 2 a a b b Gluings with twist 0, so the red "weldings" match on neighbour pants. Thomas Budzinski Minimal diameter of hyperbolic surfaces
Growth of the infinite pair of pants Hence, the tree of pants is the gluing of two copies of an infinite tree of right-angled hyperbolic hexagons: Above: infinite tree of hexagons for increasing values of a . The growth of hexagon trees corresponds to orbital counting for a subgroup of PSL 2 ( R ) generated by reflexions. This is well understood by geometers [Patterson–Sullivan, McMullen...] Thomas Budzinski Minimal diameter of hyperbolic surfaces
Bounding the diameter As for graphs, we want to show, for any centers c 1 , c 2 : � 1 � � 1 + ε � � � d hyp ( c 1 , c 2 ) ≥ log g = o . P β a g 2 We explore the balls of radius r = 1 + ε 2 β a log g around c 1 and c 2 for the hyperbolic metric on the infinite tree of pants. As for graphs, we can bound the number of "bad" steps: the volume and boundary of B r ( c 1 ) are at least a constant times what they would be in the tree of pants. � � 1 + ε β a 1 + ε 2 , and the same is So | ∂ B r ( c 1 ) | ≥ δ exp 2 β a log g = δ g true for c 2 . As for graphs, this implies that with very high probability, there is an edge between B r ( c 1 ) and B r ( c 2 ) , so d hyp ( c 1 , c 2 ) ≤ 2 r + O ( 1 ) = 1 + ε β a log g + O ( 1 ) . Thomas Budzinski Minimal diameter of hyperbolic surfaces
Recommend
More recommend
