Some Catalan Musings p. 1. Some Catalan Musings. Richard P. Stanley

Transcription

1 Some Catalan Musings p. 1 Some Catalan Musings Richard P. Stanley

2 Some Catalan Musings p. 2 An OEIS entry A000108: 1,1,2,5,14,42,132,429,...

3 Some Catalan Musings p. 2 An OEIS entry A000108: 1,1,2,5,14,42,132,429,... COMMENTS.... This is probably the longest entry in OEIS, and rightly so.

4 Some Catalan Musings p. 2 An OEIS entry A000108: 1,1,2,5,14,42,132,429,... COMMENTS.... This is probably the longest entry in OEIS, and rightly so. C n = 1 n+1( 2n n), n 0 (Catalan number)

5 Some Catalan Musings p. 3 Catalan monograph R. Stanley, Catalan Numbers, Cambridge University Press, 2015, to appear.

6 Some Catalan Musings p. 3 Catalan monograph R. Stanley, Catalan Numbers, Cambridge University Press, 2015, to appear. Includes 214 combinatorial interpretations of C n and 68 additional problems.

7 An early version (1970 s) Some Catalan Musings p. 4

8 An early version (1970 s) Some Catalan Musings p. 4

9 Some Catalan Musings p. 5

10 Some Catalan Musings p. 6 How to sample? Compare D. E. Knuth, 3:16 Bible Texts Illuminated. Sample from Bible by choosing verse 3:16 from each chapter.

11 Some Catalan Musings p. 6 How to sample? Compare D. E. Knuth, 3:16 Bible Texts Illuminated. Sample from Bible by choosing verse 3:16 from each chapter. I will be less random.

12 Some Catalan Musings p. 7 History Sharabiin Myangat, also known as Minggatu, Ming antu ( ), and Jing An (c c. 1763): a Mongolian astronomer, mathematician, and topographic scientist who worked at the Qing court in China.

13 Some Catalan Musings p. 7 History Sharabiin Myangat, also known as Minggatu, Ming antu ( ), and Jing An (c c. 1763): a Mongolian astronomer, mathematician, and topographic scientist who worked at the Qing court in China. Typical result (1730 s): sin(2α) = 2sinα n=1 C n 1 4 n 1 sin2n+1 α

14 Some Catalan Musings p. 7 History Sharabiin Myangat, also known as Minggatu, Ming antu ( ), and Jing An (c c. 1763): a Mongolian astronomer, mathematician, and topographic scientist who worked at the Qing court in China. Typical result (1730 s): sin(2α) = 2sinα n=1 C n 1 4 n 1 sin2n+1 α No combinatorics, no further work in China.

15 Some Catalan Musings p. 8 More history, via Igor Pak Euler (1751): conjectured formula for number C n of triangulations of a convex (n+2)-gon

16 Some Catalan Musings p. 9 Completion of proof Goldbach and Segner ( ): helped Euler complete the proof, in pieces. Lamé (1838): first self-contained, complete proof.

17 Some Catalan Musings p. 10 Catalan Eugène Charles Catalan (1838): wrote C n in the form and showed they counted (2n)! n!(n+1)! (nonassociative) bracketings (or parenthesizations) of a string of n + 1 letters.

18 Some Catalan Musings p. 10 Catalan Eugène Charles Catalan (1838): wrote C n in the form and showed they counted (2n)! n!(n+1)! (nonassociative) bracketings (or parenthesizations) of a string of n + 1 letters. Born in 1814 in Bruges (now in Belgium, then under Dutch rule). Studied in France and worked in France and Liège, Belgium. Died in Liège in 1894.

19 Some Catalan Musings p. 11 Why Catalan numbers? Riordan (1948): introduced the term Catalan number in Math Reviews.

20 Some Catalan Musings p. 11 Why Catalan numbers? Riordan (1948): introduced the term Catalan number in Math Reviews. Riordan (1964): used the term again in Math. Reviews.

21 Some Catalan Musings p. 11 Why Catalan numbers? Riordan (1948): introduced the term Catalan number in Math Reviews. Riordan (1964): used the term again in Math. Reviews. Riordan (1968): used the term in his book Combinatorial Identities. Finally caught on.

22 Some Catalan Musings p. 11 Why Catalan numbers? Riordan (1948): introduced the term Catalan number in Math Reviews. Riordan (1964): used the term again in Math. Reviews. Riordan (1968): used the term in his book Combinatorial Identities. Finally caught on. Gardner (1976): used the term in his Mathematical Games column in Scientific American. Real popularity began.

23 Some Catalan Musings p. 12 The primary recurrence C n+1 = n k=0 C k C n k, C 0 = 1

24 Some Catalan Musings p. 12 The primary recurrence C n+1 = n k=0 C k C n k, C 0 = 1 e

25 Some Catalan Musings p. 13 Transparent interpretations 3. Binary parenthesizations or bracketings of a string of n+1 letters (xx x)x x(xx x) (x xx)x x(x xx) xx xx

26 Some Catalan Musings p. 13 Transparent interpretations 3. Binary parenthesizations or bracketings of a string of n+1 letters (xx x)x x(xx x) (x xx)x x(x xx) xx xx ((x(xx))x)(x(xx)(xx))

27 Some Catalan Musings p. 13 Transparent interpretations 3. Binary parenthesizations or bracketings of a string of n+1 letters (xx x)x x(xx x) (x xx)x x(x xx) xx xx ((x(xx))x)(x(xx)(xx))

28 Some Catalan Musings p. 14 Binary trees 4. Binary trees with n vertices

29 Some Catalan Musings p. 14 Binary trees 4. Binary trees with n vertices

30 Some Catalan Musings p. 14 Binary trees 4. Binary trees with n vertices

31 Some Catalan Musings p. 15 Plane trees Plane tree: subtrees of a vertex are linearly ordered 6. Plane trees with n + 1 vertices

32 Plane tree recurrence Some Catalan Musings p. 16

33 Plane tree recurrence Some Catalan Musings p. 16

34 Some Catalan Musings p. 17 The natural bijection a a b c a b c d b d e f g d e f g h e c h i j h i j i f j g

35 Some Catalan Musings p. 18 Dyck paths 25. Dyck paths of length 2n, i.e., lattice paths from (0, 0) to (2n, 0) with steps (1, 1) and (1, 1), never falling below the x-axis

36 Some Catalan Musings p. 18 Dyck paths 25. Dyck paths of length 2n, i.e., lattice paths from (0, 0) to (2n, 0) with steps (1, 1) and (1, 1), never falling below the x-axis

37 Some Catalan Musings p. 18 Dyck paths 25. Dyck paths of length 2n, i.e., lattice paths from (0, 0) to (2n, 0) with steps (1, 1) and (1, 1), never falling below the x-axis

38 Some Catalan Musings p avoiding permutations 116. Permutations a 1 a 2 a n of 1,2,...,n for which there does not exist i < j < k and a j < a k < a i (called 312-avoiding) permutations)

39 Some Catalan Musings p avoiding permutations 116. Permutations a 1 a 2 a n of 1,2,...,n for which there does not exist i < j < k and a j < a k < a i (called 312-avoiding) permutations)

40 Some Catalan Musings p avoiding permutations 116. Permutations a 1 a 2 a n of 1,2,...,n for which there does not exist i < j < k and a j < a k < a i (called 312-avoiding) permutations)

41 Some Catalan Musings p. 20 Less transparent interpretations 159. Noncrossing partitions of 1, 2,..., n, i.e., partitions π = {B 1,...,B k } Π n such that if a < b < c < d and a,c B i and b,d B j, then i = j

42 Bijection with plane trees Some Catalan Musings p. 21

43 Some Catalan Musings p. 21 Bijection with plane trees

44 Some Catalan Musings p. 21 Bijection with plane trees Children of nonleaf vertices: {1,5,6}, {2},{3,4},{7,9},{8},{10,11,12}

45 Some Catalan Musings p. 22 Noncrossing partition recurrence

46 Some Catalan Musings p. 22 Noncrossing partition recurrence

47 Some Catalan Musings p avoiding permutations 115. Permutations a 1 a 2 a n of 1,2,...,n with longest decreasing subsequence of length at most two (i.e., there does not exist i < j < k, a i > a j > a k ), called 321-avoiding permutations

48 Some Catalan Musings p. 24 Bijection with Dyck paths w =

49 Some Catalan Musings p. 24 Bijection with Dyck paths w =

50 Some Catalan Musings p. 24 Bijection with Dyck paths w =

51 Some Catalan Musings p. 25 Semiorders (finite) semiorder or unit interval order: a finite subset P of R with the partial order: x < P y x < R y 1 Equivalently, no induced (3 + 1) or (2 + 2)

52 Some Catalan Musings p. 26 Semiorders (cont.) 180. Nonisomorphic n-element posets with no induced subposet isomorphic to or 3 + 1

53 Some Catalan Musings p. 27 Semiorders and Dyck paths

54 Some Catalan Musings p. 27 Semiorders and Dyck paths x x x x x x x x x x

55 Some Catalan Musings p. 27 Semiorders and Dyck paths x x x x x x x x x x

56 Some Catalan Musings p. 28 Noncrossing matchings 61. Noncrossing (complete) matchings on 2n vertices, i.e., ways of connecting 2n points in the plane lying on a horizontal line by n nonintersecting arcs, each arc connecting two of the points and lying above the points

57 Bijection to ballot sequences Some Catalan Musings p. 29

58 Some Catalan Musings p. 29 Bijection to ballot sequences left endpoint: 1 right endpoint: 1

59 Some Catalan Musings p. 30 Inverse bijection

60 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

61 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

62 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

63 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

64 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

65 Some Catalan Musings p. 30 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with leftmost available 1.

66 Some Catalan Musings p. 31 Nonnesting matchings 64. Nonnesting matchings on [2n], i.e., ways of connecting 2n points in the plane lying on a horizontal line by n arcs, each arc connecting two of the points and lying above the points, such that no arc is contained entirely below another

67 Bijection to ballot sequences Some Catalan Musings p. 32

68 Some Catalan Musings p. 32 Bijection to ballot sequences left endpoint: 1 right endpoint: 1

69 Some Catalan Musings p. 32 Bijection to ballot sequences left endpoint: 1 right endpoint: 1 Same rule as for noncrossing matchings!

70 Some Catalan Musings p. 33 Inverse bijection

71 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

72 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

73 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

74 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

75 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

76 Some Catalan Musings p. 33 Inverse bijection Scan ballot sequence from right-to-left. Connect each 1 with rightmost available 1.

77 Some Catalan Musings p. 34 Many interpretations By changing the connection rule from the 1 s to 1 s, we get infinitely many combinatorial interpretations of Catalan numbers in terms of complete matchings. All have the same bijection rule from the matchings to the ballot sequences!

78 Some Catalan Musings p. 34 Many interpretations By changing the connection rule from the 1 s to 1 s, we get infinitely many combinatorial interpretations of Catalan numbers in terms of complete matchings. All have the same bijection rule from the matchings to the ballot sequences!

79 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

80 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

81 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

82 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

83 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

84 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors

85 Some Catalan Musings p. 35 Unexpected interpretations 92. n-tuples (a 1,a 2,...,a n ) of integers a i 2 such that in the sequence 1a 1 a 2 a n 1, each a i divides the sum of its two neighbors UDUUDDUD

86 Some Catalan Musings p. 36 Cores hook lengths of a partition λ p-core: a partition with no hook lengths equal to (equivalently, divisible by) p (p,q)-core: a partition that is simultaneously a p-core and q-core

87 Some Catalan Musings p. 37 (n,n + 1)-cores 112. Integer partitions that are both n-cores and (n+1)-cores

88 Some Catalan Musings p. 38 Constructing (5,6)-cores 6 5

89 Some Catalan Musings p. 38 Constructing (5,6)-cores

90 Some Catalan Musings p. 38 Constructing (5,6)-cores

91 Some Catalan Musings p. 38 Constructing (5,6)-cores

92 Some Catalan Musings p. 38 Constructing (5,6)-cores

93 Some Catalan Musings p. 39 (4,3,1,1,1,1) is a (5,6)-core

94 Some Catalan Musings p. 40 Inversions of permutations inversion of a 1 a 2 a n S n : (a i,a j ) such that i < j, a i > a j

95 Some Catalan Musings p. 40 Inversions of permutations inversion of a 1 a 2 a n S n : (a i,a j ) such that i < j, a i > a j 186. Sets S of n non-identity permutations in S n+1 such that every pair (i,j) with 1 i < j n is an inversion of exactly one permutation in S {1243, 2134, 3412},{1324, 2314, 4123},{2134, 3124, 4123} {1324, 1423, 2341},{1243, 1342, 2341}

96 Some Catalan Musings p. 40 Inversions of permutations inversion of a 1 a 2 a n S n : (a i,a j ) such that i < j, a i > a j 186. Sets S of n non-identity permutations in S n+1 such that every pair (i,j) with 1 i < j n is an inversion of exactly one permutation in S {1243, 2134, 3412},{1324, 2314, 4123},{2134, 3124, 4123} {1324, 1423, 2341},{1243, 1342, 2341} due to R. Dewji, I. Dimitrov, A. McCabe, M. Roth, D. Wehlau, J. Wilson

97 Some Catalan Musings p. 41 A8. Algebraic interpretations (a) Number of two-sided ideals of the algebra of all (n 1) (n 1) upper triangular matrices over a field

98 Some Catalan Musings p. 41 A8. Algebraic interpretations (a) Number of two-sided ideals of the algebra of all (n 1) (n 1) upper triangular matrices over a field 0 * * * * * * * * * * * * * * * * * * * * *

99 Some Catalan Musings p. 42 Quasisymmetric functions Quasisymmetric function: a polynomial f Q[x 1,...,x n ] such that if i 1 < < i n then [x a 1 i 1 x a n i n ]f = [x a 1 1 xa n n ]f.

100 Some Catalan Musings p. 42 Quasisymmetric functions Quasisymmetric function: a polynomial f Q[x 1,...,x n ] such that if i 1 < < i n then [x a 1 i 1 x a n i n ]f = [x a 1 1 xa n n ]f. (k) Dimension (as a Q-vector space) of the ring Q[x 1,...,x n ]/Q n, where Q n denotes the ideal of Q[x 1,...,x n ] generated by all quasisymmetric functions in the variables x 1,...,x n with 0 constant term

101 Some Catalan Musings p. 42 Quasisymmetric functions Quasisymmetric function: a polynomial f Q[x 1,...,x n ] such that if i 1 < < i n then [x a 1 i 1 x a n i n ]f = [x a 1 1 xa n n ]f. (k) Dimension (as a Q-vector space) of the ring Q[x 1,...,x n ]/Q n, where Q n denotes the ideal of Q[x 1,...,x n ] generated by all quasisymmetric functions in the variables x 1,...,x n with 0 constant term Difficult proof by J.-C. Aval, F. Bergeron and N. Bergeron, 2004.

102 Some Catalan Musings p. 43 Diagonal harmonics (i) Let the symmetric group S n act on the polynomial ring A = C[x 1,...,x n,y 1,...,y n ] by w f(x 1,...,x n,y 1,...,y n ) = f(x w(1),...,x w(n),y w(1),...,y w(n) ) for all w S n. Let I be the ideal generated by all invariants of positive degree, i.e., I = f A : w f = f for all w S n, and f(0) = 0.

103 Some Catalan Musings p. 44 Diagonal harmonics (cont.) Then C n is the dimension of the subspace of A/I affording the sign representation, i.e., C n = dim{f A/I : w f = (sgnw)f for all w S n }.

104 Some Catalan Musings p. 44 Diagonal harmonics (cont.) Then C n is the dimension of the subspace of A/I affording the sign representation, i.e., C n = dim{f A/I : w f = (sgnw)f for all w S n }. Very deep proof by M. Haiman, 1994.

105 Some Catalan Musings p. 45 Generalizations & refinements A12. k-triangulation of n-gon: maximal collections of diagonals such that no k +1 of them pairwise intersect in their interiors k = 1: an ordinary triangulation superfluous edge: an edge between vertices at most k steps apart (along the boundary of the n-gon). They appear in all k-triangulations and are irrelevant.

106 Some Catalan Musings p. 46 An example Example. 2-triangulations of a hexagon (superfluous edges omitted):

107 Some Catalan Musings p. 47 Some theorems Theorem (Nakamigawa, Dress-Koolen-Moulton). All k-triangulations of an n-gon have k(n 2k 1) nonsuperfluous edges.

108 Some Catalan Musings p. 47 Some theorems Theorem (Nakamigawa, Dress-Koolen-Moulton). All k-triangulations of an n-gon have k(n 2k 1) nonsuperfluous edges. Theorem (Jonsson, Serrano-Stump). The number T k (n) of k-triangulations of an n-gon is given by T k (n) = det[c n i j ] k i,j=1 2k +i+j 1 =. i+j 1 1 i<j n 2k

109 Some Catalan Musings p. 48 Representation theory? Note. The number T k (n) is the dimension of an irreducible representation of the symplectic group Sp(2n 4).

110 Some Catalan Musings p. 48 Representation theory? Note. The number T k (n) is the dimension of an irreducible representation of the symplectic group Sp(2n 4). Is there a direct connection?

111 Some Catalan Musings p. 49 Number theory A61. Let b(n) denote the number of 1 s in the binary expansion of n. Using Kummer s theorem on binomial coefficients modulo a prime power, show that the exponent of the largest power of 2 dividing C n is equal to b(n+1) 1.

112 Some Catalan Musings p. 50 Sums of three squares Let f(n) denote the number of integers 1 k n such that k is the sum of three squares (of nonnegative integers). Well-known: lim n f(n) n = 5 6.

113 Sums of three squares Let f(n) denote the number of integers 1 k n such that k is the sum of three squares (of nonnegative integers). Well-known: lim n f(n) n = 5 6. Let g(n) denote the number of integers 1 k n such that C k is the sum of three squares. Then lim n g(n) n =??. Some Catalan Musings p. 50

114 Sums of three squares Let f(n) denote the number of integers 1 k n such that k is the sum of three squares (of nonnegative integers). Well-known: lim n f(n) n = 5 6. A63. Let g(n) denote the number of integers 1 k n such that C k is the sum of three squares. Then lim n g(n) n = 7 8. Some Catalan Musings p. 50

115 Some Catalan Musings p. 51 Analysis A65.(b) n 0 1 C n =??

116 Some Catalan Musings p. 51 Analysis A65.(b) n 0 1 = π C n 27.

117 Some Catalan Musings p. 52 Why? A65.(a) n 0 x n = 2(x+8) C n (4 x) + 24 ( xsin 1 1 ) 2 x. 2 (4 x) 5/2

118 Some Catalan Musings p. 52 Why? A65.(a) n 0 x n = 2(x+8) C n (4 x) + 24 ( xsin 1 1 ) 2 x. 2 (4 x) 5/2 Consequence of 2 ( sin 1 x ) 2 = 2 n 1 x 2n n 2( 2n n ).

119 Some Catalan Musings p. 52 Why? A65.(a) n 0 x n = 2(x+8) C n (4 x) + 24 ( xsin 1 1 ) 2 x. 2 (4 x) 5/2 Consequence of 2 ( sin 1 x ) 2 = 2 n 1 n 0 4 3n C n = 2. x 2n n 2( 2n n ).

120 An outlier Euler (1737): e = Convergents: 1,3, 19 7, ,... Some Catalan Musings p. 53

121 Some Catalan Musings p. 54 A curious generating function a n : numerator of the nth convergent a 1 = 1, a 2 = 3, a 3 = 19, a 4 = 193

122 Some Catalan Musings p. 54 A curious generating function a n : numerator of the nth convergent a 1 = 1, a 2 = 3, a 3 = 19, a 4 = n 1 a n x n n! = exp m 0 C m x m+1

123 The last slide Some Catalan Musings p. 55

124 The last slide Some Catalan Musings p. 55

125 The last slide Some Catalan Musings p. 55

126 The last slide Some Catalan Musings p. 55