a new combinatorial identity for unicellular maps via a
play

A new combinatorial identity for unicellular maps, via a direct - PowerPoint PPT Presentation

A new combinatorial identity for unicellular maps, via a direct bijective approach Guillaume Chapuy, Ecole Polytechnique (France) 13 9 17 3 FPSAC, Hagenberg, 2009. 11 Unicellular maps as polygon gluings We start with a 2 n -gon, and


  1. A new combinatorial identity for unicellular maps, via a direct bijective approach Guillaume Chapuy, ´ Ecole Polytechnique (France) 13 9 17 3 FPSAC, Hagenberg, 2009. 11

  2. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface.

  3. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface.

  4. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface.

  5. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface.

  6. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface. The image of the polygon forms the drawing of an n -edge graph on the surface. Euler’s formula relates the number of vertices to the genus of the surface : v = n + 1 − 2 g

  7. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface. 1 vertex, genus 1 3 vertices, genus 0 The image of the polygon forms the drawing of an n -edge graph on the surface. Euler’s formula relates the number of vertices to the genus of the surface : v = n + 1 − 2 g

  8. Unicellular maps as polygon gluings We start with a 2 n -gon, and we paste the edges pairwise in order to form an orientable surface. 1 vertex, genus 1 3 vertices, genus 0 The image of the polygon forms the drawing of an n -edge graph on the surface. Euler’s formula relates the number of vertices to the genus of the surface : v = n + 1 − 2 g

  9. Counting The number of unicellular maps with n edges is equal to the number of distinct matchings of the edges : (2 n )! 2 n n ! . Aim: count unicellular maps of fixed genus.

  10. Counting The number of unicellular maps with n edges is equal to the number of distinct matchings of the edges : (2 n )! 2 n n ! . Aim: count unicellular maps of fixed genus. For instance, in the planar case... Unicellular maps are exactly plane trees. Therefore the number of n -edge unicellular maps of genus 0 is : � 2 n � 1 ǫ 0 ( n ) = Cat( n ) = n + 1 n

  11. Higher genus ? For each g the number of n -edge unicellular maps of genus g has the (beautiful) form : ǫ g ( n ) = (some polynomial) × Cat( n ) For instance : ǫ 1 ( n ) = ( n +1) n ( n − 1) Cat( n ) 12 ǫ 2 ( n ) = ( n +1) n ( n − 1)( n − 2)( n − 3)(5 n − 2) Cat( n ) 1440 References : Lehman and Walsh 72 (formal power series), Harer and Zagier 86 (matrix integrals).

  12. Higher genus ? For each g the number of n -edge unicellular maps of genus g has the (beautiful) form : ǫ g ( n ) = (some polynomial) × Cat( n ) For instance : ǫ 1 ( n ) = ( n +1) n ( n − 1) Cat( n ) 12 ǫ 2 ( n ) = ( n +1) n ( n − 1)( n − 2)( n − 3)(5 n − 2) Cat( n ) 1440 References : Lehman and Walsh 72 (formal power series), Harer and Zagier 86 (matrix integrals). No combinatorial interpretation !

  13. Higher genus ? For each g the number of n -edge unicellular maps of genus g has the (beautiful) form : ǫ g ( n ) = (some polynomial) × Cat( n ) For instance : ǫ 1 ( n ) = ( n +1) n ( n − 1) Cat( n ) 12 ǫ 2 ( n ) = ( n +1) n ( n − 1)( n − 2)( n − 3)(5 n − 2) Cat( n ) 1440 References : Lehman and Walsh 72 (formal power series), Harer and Zagier 86 (matrix integrals). No combinatorial interpretation ! Note for experts: the Goulden-Nica bijection does not solve the same problem (it solves a ”Poissonized” version of the problem).

  14. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex.

  15. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  16. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  17. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  18. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  19. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  20. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges around each vertex. We started from one single polygon ⇒ the graph has only one border

  21. Map = graph + rotation system = All the information is contained in the pair formed by the graph and the cyclic ordering of edges aroud each vertex. We started from one single polygon ⇒ the graph has only one border To do: cut the 2g independant cycles of this graph in order to obtain a tree. Problem: where to cut ?

  22. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n .

  23. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n . 3 1 2 border

  24. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n . 5 6 7 8 9 4 3 1 10 2 border

  25. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n . 12 19 15 18 16 5 14 6 7 8 13 9 4 17 20 3 1 11 10 2 border

  26. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n . 12 19 15 18 16 5 14 6 13 7 9 8 17 13 3 9 11 4 17 20 3 1 11 10 2 border We compare the two natural orderings of corners around one vertex: this gives a diagram.

  27. Numbering the corners. We follow the border of the map starting from the root, and we number the corners from 1 to 2 n . 20 12 19 15 18 16 5 14 6 13 7 9 8 17 13 3 9 11 4 17 20 3 1 11 10 2 border . We compare the two natural orderings of corners . . around one vertex: this gives a diagram. 2 1

  28. Planar case In the planar case, the border-numbering and the cyclic ordering always coincide:

  29. Planar case In the planar case, the border-numbering and the cyclic ordering always coincide: For each vertex, the diagram is increasing: 3 rd 2 nd 4 th 1 st

  30. Planar case In the planar case, the border-numbering and the cyclic ordering always coincide: For each vertex, the diagram is increasing: 3 rd 2 nd 20 4 th 1 st trisection trisection Higher genus Around each vertex, a decrease in the diagram is called a trisection. . . 13 . 9 17 3 2 11 1

  31. The trisection lemma A unicellular map of genus g always has exactly 2 g trisections. Proof: simple counting argument. → It is an equivalent problem to count unicellular maps with a distinguished trisection.

  32. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  33. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  34. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  35. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  36. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  37. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1

  38. How to build a trisection : first method. - Start with a map of genus ( g − 1) with three marked vertices. - Let a 1 < a 2 < a 3 be the labels of their minimal corners. - Glue these three corners together as follows : a 3 a 2 a 1 - The resulting map has only one border : 1 → 2 → . . . . . . . . . . . . → 2 n → a 1 → → a 2 → → a 3 →

Recommend


More recommend