A sextuple equidistribution arising in Pattern Avoidance Zhicong - - PowerPoint PPT Presentation

a sextuple equidistribution arising in pattern avoidance
SMART_READER_LITE
LIVE PREVIEW

A sextuple equidistribution arising in Pattern Avoidance Zhicong - - PowerPoint PPT Presentation

A sextuple equidistribution arising in Pattern Avoidance Zhicong Lin NIMS & Jimei University 78th S eminaire Lotharingien de Combinatoire March 29, 2017 Joint work with Dongsu Kim Zhicong Lin A sextuple equidistribution arising in


slide-1
SLIDE 1

A sextuple equidistribution arising in Pattern Avoidance

Zhicong Lin

NIMS & Jimei University

78th S´ eminaire Lotharingien de Combinatoire March 29, 2017 Joint work with Dongsu Kim

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-2
SLIDE 2

Eulerian polynomials

Definition The Eulerian polynomial An(t) may be defined by Euler’s basic formula (Leonhard Euler 1755):

  • k≥0

(k + 1)ntk = An(t) (1 − t)n+1 . A1(t) = 1 A2(t) = 1 + t A3(t) = 1 + 4t + t2 A4(t) = 1 + 11t + 11t2 + t3 A5(t) = 1 + 26t + 66t2 + 26t3 + t4

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-3
SLIDE 3

Permutation Statistics

Sn: Set of permutations of [n] := {1, 2, · · · , n} Definition For π = π1π2 · · · πn ∈ Sn: DES(π) := {i ∈ [n − 1] : πi > πi+1} des(π) := |DES(π)| (Descent number). DES(3.15.24) = {1, 3}

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-4
SLIDE 4

Permutation Statistics

Sn: Set of permutations of [n] := {1, 2, · · · , n} Definition For π = π1π2 · · · πn ∈ Sn: DES(π) := {i ∈ [n − 1] : πi > πi+1} des(π) := |DES(π)| (Descent number). DES(3.15.24) = {1, 3} Theorem (Riordan 1958) An(t) =

  • π∈Sn

tdes(π).

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-5
SLIDE 5

Inversion sequences

Inversion sequences: In = {(e1, e2, . . . , en) ∈ Zn : 0 ≤ ei < i} I3 = {(0, 0, 0), (0, 0, 1), (0, 0, 2), (0, 1, 0), (0, 1, 1), (0, 1, 2)} Definition For e = (e1, e2, · · · , en) ∈ In: ASC(e) := {i ∈ [n − 1] : ei < ei+1} asc(e) := |ASC(e)| (Ascent number). ASC(0, 1, 1, 2, 0) = {1, 3}

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-6
SLIDE 6

A natural bijection: inv-code

|Sn| = |In| = n! and more...

  • π∈Sn

tdes(π) =

  • e∈In

tasc(e)

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-7
SLIDE 7

A natural bijection: inv-code

|Sn| = |In| = n! and more...

  • π∈Sn

tdes(π) =

  • e∈In

tasc(e) A natural bijection (inv-code) φ : Sn → In with φ(π) = (e1, . . . , en), where ei = |{j : j < i and πj > πi}|.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-8
SLIDE 8

A natural bijection: inv-code

|Sn| = |In| = n! and more...

  • π∈Sn

tdes(π) =

  • e∈In

tasc(e) A natural bijection (inv-code) φ : Sn → In with φ(π) = (e1, . . . , en), where ei = |{j : j < i and πj > πi}|. This proves even more:

  • π∈Sn

tDES(π) =

  • e∈In

tASC(e), where t{i1,...,ik} := ti1 · · · tik.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-9
SLIDE 9

Double Eulerian statistics

dist(e): number of distinct positive entries in e Theorem (Dumont 1974)

  • π∈Sn

tdes(π) =

  • e∈In

tdist(e).

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-10
SLIDE 10

Double Eulerian statistics

dist(e): number of distinct positive entries in e Theorem (Dumont 1974)

  • π∈Sn

tdes(π) =

  • e∈In

tdist(e). Via V-code and S-code: Theorem (Foata 1977)

  • π∈Sn

sdes(π−1)tDES(π) =

  • e∈In

sdist(e)tASC(e). Rediscovered by Visontai (2013) An essentially different proof by Aas in PP 2013 (Paris)

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-11
SLIDE 11

Gessel’s γ-positivity conjecture

Double Eulerian polynomials (Carlitz-Roselle-Scoville 1966): An(s, t) :=

  • π∈Sn

sdes(π−1)tdes(π). Conjectured by Gessel (2005): Theorem (L. 2015) The integers γn,i,j are nonnegative in: An(s, t) =

  • i,j≥0

j+2i≤n−1

γn,i,j(st)i(1 + st)j(s + t)n−1−j−2i.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-12
SLIDE 12

Permutations without double descents

✲ ✲ ✛ ✲ ✛

  • −∞

r

3

r

4

r

8

❅ ❅ ❅ ❅ ❅ ❅r2 r6

  • r7

r

5

❅ ❅ ❅ ❅ ❅ ❅ ❅

−∞

r1 Figure : Foata-Strehl actions on 34862571

NDDn: set of all permutations in Sn without double descents Theorem (Foata & Sch¨ utzenberger 1970) An(t) =

⌊(n−1)/2⌋

  • i=0

γn,iti(1 + t)n+1−2i, where γn,i = #{π ∈ NDDn : des(π) = i}. Problem

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-13
SLIDE 13

Permutations without double descents

NDDn: set of all permutations in Sn without double descents Theorem (Foata & Sch¨ utzenberger 1970) An(t) =

⌊(n−1)/2⌋

  • i=0

γn,iti(1 + t)n+1−2i, where γn,i = #{π ∈ NDDn : des(π) = i}. Problem Is there any combinatorial interpretation for γn,i,j?

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-14
SLIDE 14

Separable permutations

Restrict to the terms without s + t: π = 2413

r r r r ❇ ❇ ❇ ❇ ❇ ❍❍ ❍ ❍❍ ❍

des(π) = 1 des(π−1) = 2 First des(π) = des(π−1) Definition Permutations that avoid both the patterns 2413 and 3142 are separable permutations. West (1995): |Sn(2413, 3142)| = Sn, the nth Large Schr¨

  • der

numbers.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-15
SLIDE 15

Separable permutations

⇐ ⇒ bij. Separable permutations “di-sk” trees ⊖ ⊖ ⊕ ⊖ ⊖ ⊕ ⊖ ⊕

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-16
SLIDE 16

Descent polynomial on Separable permutations

Via combinatorial approach using “di-sk” trees: Theorem (Fu-L.-Zeng 2015)

  • π∈Sn(2413,3142)

tdes(π) =

⌊(n−1)/2⌋

  • k=0

γS

n,ktk(1 + t)n−1−2k,

where γS

n,k = |{π ∈ Sn(3142, 2413) ∩ NDDn : des(π) = k}|.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-17
SLIDE 17

021-avoiding inversion sequences

021-avoiding ⇔ positive entries are weakly increasing Via bijections with “di-sk” trees: Theorem (Fu-L.-Zeng & Corteel et al. 2015)

  • π∈Sn(2413,3142)

tdes(π) =

  • e∈In(021)

tasc(e). Problem

  • e∈In(021)

tasc(e) =

⌊(n−1)/2⌋

  • k=0

γS

n,ktk(1 + t)n−1−2k

What is the combinatorial interpretation of γS

n,k in terms of

021-avoiding inversion sequences?

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-18
SLIDE 18

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-19
SLIDE 19

Double Eulerian equidistribution

Theorem (Foata 1977)

  • π∈Sn

sdes(π−1)tDES(π) =

  • e∈In

sdist(e)tASC(e). Restricted version of Foata’s 1977 result: Theorem (Kim-L. 2016)

  • π∈Sn(2413,4213)

sdes(π−1)tDES(π) =

  • e∈In(021)

sdist(e)tASC(e). Neither Foata’s original bijection nor Aas’ approach could be applied to prove this restricted version.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-20
SLIDE 20

First application

✲ ✲ ✛ ✲ ✛

  • −∞

r

3

r

4

r

8

❅ ❅ ❅ ❅ ❅ ❅r2 r6

  • r7

r

5

❅ ❅ ❅ ❅ ❅ ❅ ❅

−∞

r1

As Sn(2413, 4213) is invariant under Foata-Strehl action: Corollary

  • e∈In(021)

tasc(e) =

⌊(n−1)/2⌋

  • k=0

γS

n,ktk(1 + t)n−1−2k,

where γS

n,k = |{e ∈ In(021) : e has no double ascents, asc(e) = k}|.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-21
SLIDE 21

Second application

0 1 0 1 2 0 4 → d

Figure : The outline of an inversion sequence

S = S(s, t; z) :=

  • n≥1

zn

  • π∈Sn(2413,4213)

sdes(π−1)tdes(π) Theorem (Double Eulerian distribution) S = t(z(s − 1) + 1)S + tz(2s − 1)S2 + z(ts + 1)S + z.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-22
SLIDE 22

Ascents on Schr¨

  • der paths

A Schr¨

  • der n-path is a lattice path on the plane from (0, 0) to

(2n, 0), never going below x-axis, using the steps (1, 1) (1, −1) (2, 0).

✉ ❅ ❅ ❅✉ ✉ ❅ ❅ ❅✉

✉ ❅ ❅ ❅✉

Corollary (Conjecture of Corteel et al. 2015) An ascent in a Schr¨

  • der path is a maximal string of consecutive up
  • steps. Denoted by SPn the set of Schr¨
  • der n-path and by asc(p)

the number of ascents of p. Then,

  • e∈In(021)

sdist(e) =

  • p∈SPn−1

sasc(p).

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-23
SLIDE 23

A sextuple equidistribution (Statistics)

For each π ∈ Sn: VID(π) := {2 ≤ i ≤ n : πi appears to the right of (πi + 1)}, the values of inverse descents of π; LMA(π) := {i ∈ [n] : πi > πj for all 1 ≤ j < i}, the positions

  • f left-to-right maxima of π;

LMI(π) := {i ∈ [n] : πi < πj for all 1 ≤ j < i}, the positions

  • f left-to-right minima of π;

RMA(π) := {i ∈ [n] : πi > πj for all j ≥ i}, the positions of right-to-left maxima of π; RMI(π) := {i ∈ [n] : πi < πj for all j ≥ i}, the positions of right-to-left minima of π;

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-24
SLIDE 24

A sextuple equidistribution (Statistics)

and for each e ∈ In: DIST(e) := {2 ≤ i ≤ n : ei = 0 and ei = ej for all j > i}, the positions of the last occurrence of distinct positive entries of e; ZERO(e) := {i ∈ [n] : ei = 0}, the positions of zeros in e; EMA(e) := {i ∈ [n] : ei = i − 1}, the positions of the entries

  • f e that achieve the maximum;

RMI(e) := {i ∈ [n] : ei < ej for all j ≥ i}, the positions of right-to-left minima of e.

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-25
SLIDE 25

A sextuple equidistribution (Main result)

Theorem (Kim-L. 2016) There exists a bijection Ψ : In(021) → Sn(2413, 4213), which transforms the sextuple (DIST, ASC, ZERO, EMA, RMI, EXPO) to (VID, DES, LMA, LMI, RMA, RMI).

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-26
SLIDE 26

The algorithm Ψ

The labeling algorithm, where temporary variables L, H, P correspond to words label, height, position, works as follows:

1

(Start) L ← 1 (This means that 1 is assigned to L); draw the diagonal (line) y = x on d(e) and label the highest east step touched by the diagonal, say Ek, with L; L ← L + 1, H ← dk, P ← k; go to (2), if EP is a red east step (i.e. k = 1), otherwise go to (3);

2

draw the leftmost new line that touches at least one unlabeled black east step or a labelable red east step; label the highest east step touched by this new line, say Ek, with L; L ← L + 1, H ← dk, P ← k; go to (2), if EP is a red east step, otherwise go to (3);

3

go to (5), if there is a black east step Ej with j > P and height dj = H,

  • therwise go to (4);

4

move from EP along the two-colored Dyck path d(e) to the left and along the lines that were already drawn to the southwest until we arrive at the first unlabeled east step that is a black step or a labelable red step, say Ek; label Ek with L; L ← L + 1, H ← dk, P ← k; go to (2), if EP is a red east step, otherwise go to (3);

5

draw the leftmost line beginning at an east step right to EP which touches at least one black east step; label the highest east step touched by this new line, say Ek, with L; L ← L + 1, H ← dk, P ← k; go to (3).

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-27
SLIDE 27

Flowchart of Ψ

1 2 3 4 5 T

last not last b r r b b L = n + 1 r L = n + 1 L = n + 1 L = n + 1

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-28
SLIDE 28

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-29
SLIDE 29

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-30
SLIDE 30

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-31
SLIDE 31

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-32
SLIDE 32

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-33
SLIDE 33

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-34
SLIDE 34

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-35
SLIDE 35

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7 8

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-36
SLIDE 36

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7 8 l6 9

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-37
SLIDE 37

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7 8 l6 9 l7 10

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-38
SLIDE 38

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7 8 l6 9 l7 10 11

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-39
SLIDE 39

The algorithm Ψ (An example)

→ Ψ 0 1 0 0 1 3 0 7 0 0 7 10 l1 1 l2 2 3 l3 4 5 l4 6 l5 7 8 l6 9 l7 10 11 l8 12

Figure : An example of the algorithm Ψ

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance

slide-40
SLIDE 40

Merci pour votre attention

Zhicong Lin A sextuple equidistribution arising in Pattern Avoidance