invariants de tutte et triangulations avec mod ele d ising
play

Invariants de Tutte et triangulations avec mod` ele dIsing Marie - PowerPoint PPT Presentation

Invariants de Tutte et triangulations avec mod` ele dIsing Marie Albenque (CNRS and LIX) joint work with Laurent M enard (Paris Nanterre) and Gilles Schaeffer (CNRS and LIX) Journ ees ALEA, Mars 2019 I - Local limit of triangulations


  1. Invariants de Tutte et triangulations avec mod` ele d’Ising Marie Albenque (CNRS and LIX) joint work with Laurent M´ enard (Paris Nanterre) and Gilles Schaeffer (CNRS and LIX) Journ´ ees ALEA, Mars 2019

  2. I - Local limit of triangulations without matter

  3. Planar Maps as discrete planar metric spaces A triangulation is the proper embedding of a finite connected graph in the 2-dimensional sphere seen up to continuous deformations, such that all the faces have degree 3.

  4. Planar Maps as discrete planar metric spaces A triangulation is the proper embedding of a finite connected graph in the 2-dimensional sphere seen up to continuous deformations, such that all the faces have degree 3.

  5. Planar Maps as discrete planar metric spaces A triangulation is the proper embedding of a finite connected graph in the 2-dimensional sphere seen up to continuous deformations, such that all the faces have degree 3. Plane maps are rooted : by orienting an edge. Distance between two vertices = number of edges between them. Planar map = Metric space

  6. ”Classical” large random triangulations Take a triangulation with n edges uniformly at random. What does it look like if n is large ? Local point of view : Look at neighborhoods of the root

  7. ”Classical” large random triangulations Take a triangulation with n edges uniformly at random. What does it look like if n is large ? Local point of view : Look at neighborhoods of the root The local topology on finite maps is induced by the distance: d loc ( m, m ′ ) := (1 + max { r ≥ 0 : B r ( m ) = B r ( m ′ ) } ) − 1 where B r ( m ) is the graph made of all the vertices and edges of m which are within distance r from the root. Courtesy of Igor Kortchemski

  8. Local convergence: simple examples 0 1 2 n Root = 0

  9. Local convergence: simple examples − → ( Z + , 0) 0 1 2 n Root = 0

  10. Local convergence: simple examples − → ( Z + , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root

  11. Local convergence: simple examples − → ( Z + , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root

  12. Local convergence: simple examples − → ( Z + , 0) − → ( Z , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root

  13. Local convergence: simple examples − → ( Z + , 0) − → ( Z , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root − → ( Z , 0) 2 Root does not matter n 1

  14. Local convergence: simple examples − → ( Z + , 0) − → ( Z , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root − → ( Z , 0) 2 Root does not matter n 1 n Z 2 � � − → + , 0 0 n

  15. Local convergence: simple examples − → ( Z + , 0) − → ( Z , 0) 0 1 2 0 1 2 n n Root = 0 Uniformly chosen root − → ( Z , 0) 2 Root does not matter n 1 n n Z 2 Z 2 , 0 � � � � − → + , 0 − → 0 n 0 n Uniformly chosen root

  16. Local convergence: more complicated examples Uniform plane trees with n vertices: 1/2 1/5 1/2 1/5 1/5 1/5 n = 1 n = 2 n = 3 n = 4 1/5

  17. Local convergence: more complicated examples Uniform plane trees with n vertices: 1/2 1/5 1/2 1/5 1/5 1/5 n = 1 n = 2 n = 3 n = 4 1/5 n = 500 n = 1000

  18. Local convergence: more complicated examples Uniform plane trees with n vertices: 1/2 1/5 1/2 1/5 1/5 1/5 n = 1 n = 2 n = 3 n = 4 1/5 The limit is a probability distribution on infinite trees with one infinite branch. [Kesten] n = 500 n = 1000

  19. Local convergence of uniform triangulations Theorem [Angel – Schramm, ’03] As n → ∞ , the uniform distribution on triangulations of size n converges weakly to a probability measure called the Uniform Infinite Planar Triangulation (or UIPT ) for the local topology . Courtesy of Igor Kortchemski Courtesy of Timothy Budd

  20. Local convergence of uniform triangulations Theorem [Angel – Schramm, ’03] As n → ∞ , the uniform distribution on triangulations of size n converges weakly to a probability measure called the Uniform Infinite Planar Triangulation (or UIPT ) for the local topology . Some properties of the UIPT: • The UIPT has almost surely one end [Angel – Schramm, ’03] • Volume (nb. of vertices) and perimeters of balls known to some extent. E [ | B r ( T ∞ ) | ] ∼ 2 7 r 4 For example [Angel ’04, Curien – Le Gall ’12] • Simple random Walk is recurrent [Gurel-Gurevich and Nachmias ’13]

  21. Local convergence of uniform triangulations Theorem [Angel – Schramm, ’03] As n → ∞ , the uniform distribution on triangulations of size n converges weakly to a probability measure called the Uniform Infinite Planar Triangulation (or UIPT ) for the local topology . Some properties of the UIPT: • The UIPT has almost surely one end [Angel – Schramm, ’03] • Volume (nb. of vertices) and perimeters of balls known to some extent. E [ | B r ( T ∞ ) | ] ∼ 2 7 r 4 For example [Angel ’04, Curien – Le Gall ’12] • Simple random Walk is recurrent [Gurel-Gurevich and Nachmias ’13] Universality : we expect the same behavior for slightly different models (e.g. quadrangulations, triangulations without loops, ...)

  22. II - Ising model on random maps

  23. Adding matter: Ising model on triangulations First, Ising model on a finite deterministic graph: Spin configuration on G : G = ( V, E ) finite graph σ : V → {− 1 , +1 } . − + − − + −

  24. Adding matter: Ising model on triangulations First, Ising model on a finite deterministic graph: Spin configuration on G : G = ( V, E ) finite graph σ : V → {− 1 , +1 } . − Ising model on G: take a random + − spin configuration with probability − + P ( σ ) ∝ e − β � v ∼ v ′ 1 { σ ( v ) � = σ ( v ′ ) } 2 − β > 0 : inverse temperature. h = 0 : no magnetic field.

  25. Adding matter: Ising model on triangulations First, Ising model on a finite deterministic graph: Spin configuration on G : G = ( V, E ) finite graph σ : V → {− 1 , +1 } . − m ( σ ) = 4 m ( σ ) = 4 Ising model on G: take a random + − spin configuration with probability − + P ( σ ) ∝ e − β � v ∼ v ′ 1 { σ ( v ) � = σ ( v ′ ) } 2 − β > 0 : inverse temperature. h = 0 : no magnetic field. Combinatorial formulation: P ( σ ) ∝ ν m ( σ ) with m ( σ ) = number of monochromatic edges and ν = e β .

  26. Adding matter: Ising model on triangulations T n = { rooted planar triangulations with 3 n edges } . Random triangulation with spins in T n with probability ∝ ν m ( T,σ ) ?

  27. Adding matter: Ising model on triangulations T n = { rooted planar triangulations with 3 n edges } . Random triangulation with spins in T n with probability ∝ ν m ( T,σ ) ? = ν m ( T,σ ) δ | e ( T ) | =3 n � � P ν { ( T, σ ) } . n [ t 3 n ] Q ( ν, t ) where Q ( ν, t ) = generating series of Ising-weighted triangulations : � � ν m ( T,σ ) t e ( T ) . Q ( ν, t ) = T ∈T f σ : V ( T ) →{− 1 , +1 }

  28. Adding matter: Ising model on triangulations T n = { rooted planar triangulations with 3 n edges } . Random triangulation with spins in T n with probability ∝ ν m ( T,σ ) ? = ν m ( T,σ ) δ | e ( T ) | =3 n � � P ν { ( T, σ ) } . n [ t 3 n ] Q ( ν, t ) where Q ( ν, t ) = generating series of Ising-weighted triangulations : � � ν m ( T,σ ) t e ( T ) . Q ( ν, t ) = T ∈T f σ : V ( T ) →{− 1 , +1 } This is a probability distribution on triangulations with spins. But, forgetting the Remark: spins gives a probability a distribution on triangulations without spins different from the uniform distribution .

  29. Adding matter: New asymptotic behavior Counting exponent for undecorated maps: coeff [ t n ] of generating series of (undecorated) maps (e.g.: triangulations, quadrangulations, general maps, simple maps,...) ∼ κρ − n n − 5 / 2 Note : κ and ρ depend on the combinatorics of the model.

  30. Adding matter: New asymptotic behavior Counting exponent for undecorated maps: coeff [ t n ] of generating series of (undecorated) maps (e.g.: triangulations, quadrangulations, general maps, simple maps,...) ∼ κρ − n n − 5 / 2 Note : κ and ρ depend on the combinatorics of the model. Theorem [Bernardi – Bousquet-M´ elou 11] For every ν the series Q ( ν, t ) is algebraic, has ρ ν > 0 as unique dominant singularity and satisfies � 1 κ ρ − n ν c n − 7 / 3 if ν = ν c = 1 + 7 , √ [ t 3 n ] Q ( ν, t ) ∼ κ ρ − n n − 5 / 2 n →∞ if ν � = ν c . ν This suggests an unusual behavior of the underlying maps for ν = ν c . See also [Boulatov – Kazakov 1987], [Bousquet-Melou – Schaeffer 03] and [Bouttier – Di Francesco – Guitter 04].

  31. III - Results and idea of proofs

  32. Local convergence of triangulations with spins ν m ( T,σ ) � � Probability measure on triangulations P ν { ( T, σ ) } = [ t 3 n ] Q ( ν, t ) . n of T n with a spin configuration: Theorem [AMS] As n → ∞ , the sequence P ν n converges weakly to a probability measure P ν for the local topology . The measure P ν is supported on infinite triangulations with one end.

  33. Local Topology for planar maps : balls Definition: The local topology on M f is induced by the distance: d loc ( m, m ′ ) := (1 + max { r ≥ 0 : B r ( m ) = B r ( m ′ ) } ) − 1 where B r ( m ) is the graph made of all the faces of m with at least one vertex at distance r − 1 from the root.

Recommend


More recommend