Relations between the shape of a permutation and the shape of the - - PDF document

relations between the shape of a permutation and the
SMART_READER_LITE
LIVE PREVIEW

Relations between the shape of a permutation and the shape of the - - PDF document

Relations between the shape of a permutation and the shape of the base poset derived from the corresponding Lehmer codes 2013 July 5 at Permutation Pattens 2013 Masaya Tomie Morioka University tomie@morioka-u.ac.jp 1. Lehmer Codes and Weak


slide-1
SLIDE 1

Relations between the shape of a permutation and the shape of the base poset derived from the corresponding Lehmer codes 2013 July 5 at Permutation Pattens 2013 Masaya Tomie Morioka University tomie@morioka-u.ac.jp

slide-2
SLIDE 2
  • 1. Lehmer Codes and Weak Bruhat Order
  • 2. Denoncourt’s Work
  • 3. Relations Between ω and Mω
  • 4. Relations Between ∆(ω) and Mω
slide-3
SLIDE 3
  • 1. Lehmer Codes and Weak Bruhat

Order ω = ω(1)ω(2) · · · ω(n) ∈ Sn

slide-4
SLIDE 4
  • 1. Lehmer Codes and Weak Bruhat

Order ω = ω(1)ω(2) · · · ω(n) ∈ Sn c(ω) = (c1(ω), c2(ω), · · · , cn(ω)) Lehmer Code

slide-5
SLIDE 5
  • 1. Lehmer Codes and Weak Bruhat

Order ω = ω(1)ω(2) · · · ω(n) ∈ Sn c(ω) = (c1(ω), c2(ω), · · · , cn(ω)) Lehmer Code ⇐ ⇒ c1(ω) : the number of i (≥ 1) such that ω(1) > ω(i) c2(ω) : the number of i (≥ 2) such that ω(2) > ω(i) . . . cn(ω) : the number of i (≥ n) such that ω(n) > ω(i)

slide-6
SLIDE 6

Example ω = 423615

slide-7
SLIDE 7

Example ω = 423615

  • c1 = 3

423615

slide-8
SLIDE 8

Example ω = 423615

  • c1 = 3

423615

  • c2 = 1

423615

slide-9
SLIDE 9

Example ω = 423615

  • c1 = 3

423615

  • c2 = 1

423615

  • c3 = 1

423615

slide-10
SLIDE 10

Example ω = 423615

  • c1 = 3

423615

  • c2 = 1

423615

  • c3 = 1

423615

  • c4 = 2

423615

  • c5 = 0

423615

  • c6 = 0

423615 c(ω) = (3, 1, 1, 2, 0, 0)

slide-11
SLIDE 11

Example ω = 423615

  • c1 = 3

423615

  • c2 = 1

423615

  • c3 = 1

423615

  • c4 = 2

423615

  • c5 = 0

423615

  • c6 = 0

423615 c(ω) = (3, 1, 1, 2, 0, 0) Endow a product order on Lehmer Codes

slide-12
SLIDE 12

ω = ω(1)ω(2) · · · ω(n) ∈ Sn Inv(ω) = {(i, j)|i < j, ω(i) > ω(j)}

slide-13
SLIDE 13

ω = ω(1)ω(2) · · · ω(n) ∈ Sn Inv(ω) = {(i, j)|i < j, ω(i) > ω(j)} Definition (Weak Bruhat Order) For ω, τ ∈ Sn ω ≤ τ ⇐ ⇒ Inv(ω) ⊂ Inv(τ)

slide-14
SLIDE 14

ω = ω(1)ω(2) · · · ω(n) ∈ Sn Inv(ω) = {(i, j)|i < j, ω(i) > ω(j)} Definition (Weak Bruhat Order) For ω, τ ∈ Sn ω ≤ τ ⇐ ⇒ Inv(ω) ⊂ Inv(τ)

123 213 132 312 231 321

slide-15
SLIDE 15
  • 3. Denoncourt’s Work
  • H. Denoncourt,

A refinement of weak order intervals into dis- tributive lattices, arXiv:1102.2689.

slide-16
SLIDE 16
  • 3. Denoncourt’s Work
  • H. Denoncourt,

A refinement of weak order intervals into dis- tributive lattices, arXiv:1102.2689. ω ∈ Sn c(ω) : The corresponding Lehmer code

slide-17
SLIDE 17
  • 3. Denoncourt’s Work
  • H. Denoncourt,

A refinement of weak order intervals into dis- tributive lattices, arXiv:1102.2689. ω ∈ Sn c(ω) : The corresponding Lehmer code Λω := [e, ω] Interval in Weak Bruhat Order

slide-18
SLIDE 18
  • 3. Denoncourt’s Work
  • H. Denoncourt,

A refinement of weak order intervals into dis- tributive lattices, arXiv:1102.2689. ω ∈ Sn c(ω) : The corresponding Lehmer code Λω := [e, ω] Interval in Weak Bruhat Order Theorem (Denoncourt 2011)

  • 1. c is an order preserving map
  • 2. c(Λω) is a distributive lattice in Nn
slide-19
SLIDE 19

Definition (Denoncourt 2011) For ω ∈ Sn, i with ci(ω) ̸= 0 and 1 ≤ x ≤ ci(ω), define mi,x(ω) ∈ Nn s.t.

slide-20
SLIDE 20

Definition (Denoncourt 2011) For ω ∈ Sn, i with ci(ω) ̸= 0 and 1 ≤ x ≤ ci(ω), define mi,x(ω) ∈ Nn s.t.

  • 1. j-th (j < i) coordinate of mi,x(ω) is 0
slide-21
SLIDE 21

Definition (Denoncourt 2011) For ω ∈ Sn, i with ci(ω) ̸= 0 and 1 ≤ x ≤ ci(ω), define mi,x(ω) ∈ Nn s.t.

  • 1. j-th (j < i) coordinate of mi,x(ω) is 0
  • 2. i-th coordinate of mi,x(ω) is x
slide-22
SLIDE 22

Definition (Denoncourt 2011) For ω ∈ Sn, i with ci(ω) ̸= 0 and 1 ≤ x ≤ ci(ω), define mi,x(ω) ∈ Nn s.t.

  • 1. j-th (j < i) coordinate of mi,x(ω) is 0
  • 2. i-th coordinate of mi,x(ω) is x
  • 3. For j > i with ω(i) > ω(j),

j-th coordinate of mi,x(ω) is 0

slide-23
SLIDE 23

Definition (Denoncourt 2011) For ω ∈ Sn, i with ci(ω) ̸= 0 and 1 ≤ x ≤ ci(ω), define mi,x(ω) ∈ Nn s.t.

  • 1. j-th (j < i) coordinate of mi,x(ω) is 0
  • 2. i-th coordinate of mi,x(ω) is x
  • 3. For j > i with ω(i) > ω(j),

j-th coordinate of mi,x(ω) is 0

  • 4. For j > i with ω(i) < ω(j),

j-th coordinate of mi,x(ω) is max{0, x − ci,j(x)} where ci,j is the number of i ≤ k ≤ j s.t. ω(i) > ω(k).

slide-24
SLIDE 24

We denote the join-irreducibles by Mω.

slide-25
SLIDE 25

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)}

slide-26
SLIDE 26

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642

slide-27
SLIDE 27

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642 c(ω) = (4, 2, 4, 0, 2, 1, 0)

slide-28
SLIDE 28

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642 c(ω) = (4, 2, 4, 0, 2, 1, 0) m1,4 = (4, , , , , , )

slide-29
SLIDE 29

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642 c(ω) = (4, 2, 4, 0, 2, 1, 0) m1,4 = (4, , , , , , ) m1,4 = (4, 0, , 0, , 0, 0) 5371642

slide-30
SLIDE 30

We denote the join-irreducibles by Mω. Theorem (Denoncourt 2011) The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642 c(ω) = (4, 2, 4, 0, 2, 1, 0) m1,4 = (4, , , , , , ) m1,4 = (4, 0, , 0, , 0, 0) 5371642 m1,4 = (4, 0, 3, 0, , 0, 0) 5371642

slide-31
SLIDE 31

We denote the join-irreducibles by Mω. Theorem The join-irreducibles of c(Λω) are Mω = {mi,x ∈ Nn|ci(ω) ̸= 0, 1 ≤ x ≤ ci(ω)} Example ω = 5371642 c(ω) = (4, 2, 4, 0, 2, 1, 0) m1,4 = (4, , , , , , ) m1,4 = (4, 0, , 0, , 0, 0) 5371642 m1,4 = (4, 0, 3, 0, , 0, 0) 5371642 m1,4 = (4, 0, 3, 0, 2, 0, 0) 5371642

slide-32
SLIDE 32

m1,4 = (4, 0, 3, 0, 2, 0, 0) m1,3 = (3, 0, 2, 0, 1, 0, 0) m1,2 = (2, 0, 1, 0, 0, 0, 0) m1,1 = (1, 0, 0, 0, 0, 0, 0) m2,2 = (0, 2, 2, 0, 1, 1, 0) m2,1 = (0, 1, 1, 0, 0, 0, 0) m3,4 = (0, 0, 4, 0, 0, 0, 0) m3,3 = (0, 0, 3, 0, 0, 0, 0) m3,2 = (0, 0, 2, 0, 0, 0, 0) m3,1 = (0, 0, 1, 0, 0, 0, 0) m5,2 = (0, 0, 0, 0, 2, 0, 0) m5,1 = (0, 0, 0, 0, 1, 0, 0) m6,1 = (0, 0, 0, 0, 0, 1, 0)

slide-33
SLIDE 33

m11 m12 m13 m14 m21 m22 m31 m32 m33 m34 m51 m52 m61

slide-34
SLIDE 34
  • 3. Relations Between ω and Mω
slide-35
SLIDE 35
  • 3. Relations Between ω and Mω

Lemma ω is a 231-avoiding permutation = ⇒ Mω is disjoint union of chains.

slide-36
SLIDE 36
  • 3. Relations Between ω and Mω

Lemma ω is a 231-avoiding permutation = ⇒ Mω is disjoint union of chains. M4231 is disjoint union of length 2 and 1 chains.

slide-37
SLIDE 37
  • 3. Relations Between ω and Mω

Lemma ω is a 231-avoiding permutation = ⇒ Mω is disjoint union of chains. M4231 is disjoint union of length 2 and 1 chains. Definition P, Q Posets P is called to be Q free poset iff there are no subposets R ⊂ P s.t. R ≃ Q.

slide-38
SLIDE 38
  • 3. Relations Between ω and Mω

Lemma ω is a 231-avoiding permutation = ⇒ Mω is disjoint union of chains. M4231 is disjoint union of length 2 and 1 chains. Definition P, Q Posets P is called to be Q free poset iff there are no subposets R ⊂ P s.t. R ≃ Q.

  • 1. (2 + 2)–(3 + 1)-free poset is enumerated by

Catalan number.

  • 2. A (2 + 2)-free poset is a interval order.
slide-39
SLIDE 39
  • 3. Relations Between ω and Mω

Lemma ω is a 231-avoiding permutation = ⇒ Mω is disjoint union of chains. M4231 is disjoint union of length 2 and 1 chains. Definition P, Q Posets P is called to be Q free poset iff there are no subposets R ⊂ P s.t. R ≃ Q.

  • 1. (2 + 2)–(3 + 1)-free poset is enumerated by

Catalan number.

  • 2. A (2 + 2)-free poset is a interval order.

A poset P is B2-free iff P has no 4 elements isomorphic to Boolean algebra of rank 2.

slide-40
SLIDE 40

Theorem (T. 2011) ω is a 3412-3421-avoiding permutation ⇐ ⇒ Mω is a B2 free poset.

  • M. Tomie,

A relation between the shape of a permuta- tion and the shape of the base poset derived from the Lehmer codes, arXiv:1111.3094.

slide-41
SLIDE 41

Theorem (T. 2011) ω is a 3412-3421-avoiding permutation ⇐ ⇒ Mω is a B2 free poset.

  • M. Tomie,

A relation between the shape of a permuta- tion and the shape of the base poset derived from the Lehmer codes, arXiv:1111.3094. 3412-3421-avoiding permutation is called Schr¨

  • der

permutation

slide-42
SLIDE 42

Theorem (T. 2011) ω is a 3412-3421-avoiding permutation ⇐ ⇒ Mω is a B2 free poset.

  • M. Tomie,

A relation between the shape of a permuta- tion and the shape of the base poset derived from the Lehmer codes, arXiv:1111.3094. 3412-3421-avoiding permutation is called Schr¨

  • der

permutation Example ω = 315462

m11 m12 m31 m32 m41 m51

slide-43
SLIDE 43

Consider the number of components of Mω

slide-44
SLIDE 44

Consider the number of components of Mω Definition For ω ∈ Sn Consider a graph G(ω) s.t.

slide-45
SLIDE 45

Consider the number of components of Mω Definition For ω ∈ Sn Consider a graph G(ω) s.t.

  • 1. vertex set {i|∃j > i, s.t.ω(i) > ω(j)}
slide-46
SLIDE 46

Consider the number of components of Mω Definition For ω ∈ Sn Consider a graph G(ω) s.t.

  • 1. vertex set {i|∃j > i, s.t.ω(i) > ω(j)}
  • 2. Connect i and j (i < j) if ∃k > j s.t.

st(ω(i)ω(j)ω(k)) = 231.

slide-47
SLIDE 47

Example ω = 6472315 c(ω) = (5, 3, 4, 1, 1, 0, 0)

slide-48
SLIDE 48

Example ω = 6472315 c(ω) = (5, 3, 4, 1, 1, 0, 0) m1,5 = (5, 0, 4, 0, 0, 0, 0) m1,4 = (4, 0, 3, 0, 0, 0, 0) m1,3 = (3, 0, 2, 0, 0, 0, 0) m1,2 = (2, 0, 1, 0, 0, 0, 0) m1,1 = (1, 0, 0, 0, 0, 0, 0) m2,3 = (0, 3, 3, 0, 0, 0, 1) m2,2 = (0, 2, 2, 0, 0, 0, 0) m2,1 = (0, 1, 1, 0, 0, 0, 0) m3,4 = (0, 0, 4, 0, 0, 0, 0) m3,3 = (0, 0, 3, 0, 0, 0, 0) m3,2 = (0, 0, 2, 0, 0, 0, 0) m3,1 = (0, 0, 1, 0, 0, 0, 0) m4,1 = (0, 0, 0, 1, 1, 0, 0) m5,1 = (0, 0, 0, 0, 1, 0, 0)

slide-49
SLIDE 49
slide-50
SLIDE 50

m11 m12 m13 m14 m21 m22 m31 m32 m33 m34 m23 m15 m41 m51

slide-51
SLIDE 51

m11 m12 m13 m14 m21 m22 m31 m32 m33 m34 m23 m15 m41 m51

1 2 3 4 5

slide-52
SLIDE 52

m11 m12 m13 m14 m21 m22 m31 m32 m33 m34 m23 m15 m41 m51

1 2 3 4 5

Theorem The number of components of Mω equals to that of G(ω)

slide-53
SLIDE 53

Problem

  • 1. When Mω becomes tree?
slide-54
SLIDE 54
  • 4. Relations Between ∆(ω) and Mω
slide-55
SLIDE 55
  • 4. Relations Between ∆(ω) and Mω

ω ∈ Sn ∆(ω) := {(i, j)|i < j, ω(i) > ω(j)}

slide-56
SLIDE 56
  • 4. Relations Between ∆(ω) and Mω

ω ∈ Sn ∆(ω) := {(i, j)|i < j, ω(i) > ω(j)} Endow a partial order on ∆(ω) (i, j) ≤ (k, l) ⇐ ⇒ i ≤ k < l ≤ j ∆(ω) becomes a poset

slide-57
SLIDE 57
  • 4. Relations Between ∆(ω) and Mω

ω ∈ Sn ∆(ω) := {(i, j)|i < j, ω(i) > ω(j)} Endow a partial order on ∆(ω) (i, j) ≤ (k, l) ⇐ ⇒ i ≤ k < l ≤ j ∆(ω) becomes a poset Lemma ♯∆(ω) = ♯Inv(ω)

slide-58
SLIDE 58

Remark (Motivation of ∆(ω)) Sn is the Weyl group of type An−1 {α1, α2, · · · , αn−1} fundamental roots

slide-59
SLIDE 59

Remark (Motivation of ∆(ω)) Sn is the Weyl group of type An−1 {α1, α2, · · · , αn−1} fundamental roots ∆ the set of roots ∆+(∆−) positive roots (negative roots) ∆ = ∆+ ⊎ ∆−

slide-60
SLIDE 60

Remark (Motivation of ∆(ω)) Sn is the Weyl group of type An−1 {α1, α2, · · · , αn−1} fundamental roots ∆ the set of roots ∆+(∆−) positive roots (negative roots) ∆ = ∆+ ⊎ ∆− Endow a partial order on ∆ s.t. α ≤ β ⇐ ⇒ β − α = ∑

ki≥0 kiαi

slide-61
SLIDE 61

Remark (Motivation of ∆(ω)) Sn is the Weyl group of type An−1 {α1, α2, · · · , αn−1} fundamental roots ∆ the set of roots ∆+(∆−) positive roots (negative roots) ∆ = ∆+ ⊎ ∆− Endow a partial order on ∆ s.t. α ≤ β ⇐ ⇒ β − α = ∑

ki≥0 kiαi

˜ ∆(ω) = {α|α ∈ ∆+, ω(α) ∈ ∆−}

slide-62
SLIDE 62

Remark (Motivation of ∆(ω)) Sn is the Weyl group of type An−1 {α1, α2, · · · , αn−1} fundamental roots ∆ the set of roots ∆+(∆−) positive roots (negative roots) ∆ = ∆+ ⊎ ∆− Endow a partial order on ∆ s.t. α ≤ β ⇐ ⇒ β − α = ∑

ki≥0 kiαi

˜ ∆(ω) = {α|α ∈ ∆+, ω(α) ∈ ∆−} It is known that ˜ ∆(ω) ≃ ∆(ω) as a poset

slide-63
SLIDE 63

We define a map Φω : Mω → ∆(ω) Φω(mi,x) := (i, jx) where (i, j1), (i, j2), · · · (i, jx), · · · ) ∈ Inv(ω) with j1 < j2 < · · · < jx < · · ·

slide-64
SLIDE 64

We define a map Φω : Mω → ∆(ω) Φω(mi,x) := (i, jx) where (i, j1), (i, j2), · · · (i, jx), · · · ) ∈ Inv(ω) with j1 < j2 < · · · < jx < · · · Proposition Φω is an order preserving bijection

slide-65
SLIDE 65

We define a map Φω : Mω → ∆(ω) Φω(mi,x) := (i, jx) where (i, j1), (i, j2), · · · (i, jx), · · · ) ∈ Inv(ω) with j1 < j2 < · · · < jx < · · · Proposition Φω is an order preserving bijection But Φω is not poset isomorphism in general

slide-66
SLIDE 66

We define a map Φω : Mω → ∆(ω) Φω(mi,x) := (i, jx) where (i, j1), (i, j2), · · · (i, jx), · · · ) ∈ Inv(ω) with j1 < j2 < · · · < jx < · · · Proposition Φω is an order preserving bijection But Φω is not poset isomorphism in general Theorem (T) Φω is a poset isomorphism if and only if ω is a 321-avoiding permutation.

slide-67
SLIDE 67

We define a map Φω : Mω → ∆(ω) Φω(mi,x) := (i, jx) where (i, j1), (i, j2), · · · (i, jx), · · · ) ∈ Inv(ω) with j1 < j2 < · · · < jx < · · · Proposition Φω is an order preserving bijection But Φω is not poset isomorphism in general Theorem (T) Φω is a poset isomorphism if and only if ω is a 321-avoiding permutation. Remark A 321-avoiding permutation is a fully com- mutative element.

slide-68
SLIDE 68

Problem Are there natural generalizations of this fact to Weyl groups ?