• 検索結果がありません。

Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approach

N/A
N/A
Protected

Academic year: 2022

シェア "Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approach"

Copied!
36
0
0

読み込み中.... (全文を見る)

全文

(1)

Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approach

Nguyen Hoang Nghia

(in collaboration with G´ erard Duchamp, Thomas Krajewski and Adrian Tanas˘ a)

Laboratoire d’Informatique de Paris Nord, Universit´ e Paris 13

arXiv:1301.0782 [math.CO]

accepted by

Discrete Mathematics and Theoretical Computer Science Proceedings

70th S´ eminaire Lotharingien de Combinatoire

March 25, 2013

(2)

Plan

1 Matroids: The Tutte polynomial and the Hopf algebra

2 Characters of the Hopf algebras of matroids

3 Convolution formula for the Tutte polynomials for matroids

4 Proof of the universality of the Tutte polynomials for matroids

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 2 / 18

(3)

Matroid theory - some definitions

Definition 1.1

A matroid M = (E, I) is a pair (E , I ) a finite set: E

a collection of subsets of E: I satisfying:

I is non-empty,

every subset of every member of I is also in I,

if X , Y ∈ I and |X | = |Y | + 1, then ∃ x ∈ X − Y s.t. Y ∪ {x} ∈ I. The set E : the ground set of the matroid

The members of I : the independent sets of the matroid.

(4)

Matroid theory - some definitions

Definition 1.1

A matroid M = (E, I) is a pair (E , I ) a finite set: E

a collection of subsets of E: I satisfying:

I is non-empty,

every subset of every member of I is also in I,

if X , Y ∈ I and |X | = |Y | + 1, then ∃ x ∈ X − Y s.t. Y ∪ {x} ∈ I. The set E : the ground set of the matroid

The members of I : the independent sets of the matroid.

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 3 / 18

(5)

Matroid theory - some definitions

Definition 1.1

A matroid M = (E, I) is a pair (E , I ) a finite set: E

a collection of subsets of E: I satisfying:

I is non-empty,

every subset of every member of I is also in I,

if X , Y ∈ I and |X | = |Y | + 1, then ∃ x ∈ X − Y s.t. Y ∪ {x} ∈ I. The set E : the ground set of the matroid

The members of I : the independent sets of the matroid.

(6)

Matroid theory - some definitions

Definition 1.1

A matroid M = (E, I) is a pair (E , I ) a finite set: E

a collection of subsets of E: I satisfying:

I is non-empty,

every subset of every member of I is also in I,

if X , Y ∈ I and |X | = |Y | + 1, then ∃ x ∈ X − Y s.t. Y ∪ {x} ∈ I.

The set E : the ground set of the matroid

The members of I : the independent sets of the matroid.

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 3 / 18

(7)

Matroid theory - some definitions

Definition 1.1

A matroid M = (E, I) is a pair (E , I ) a finite set: E

a collection of subsets of E: I satisfying:

I is non-empty,

every subset of every member of I is also in I,

if X , Y ∈ I and |X | = |Y | + 1, then ∃ x ∈ X − Y s.t. Y ∪ {x} ∈ I.

The set E : the ground set of the matroid

The members of I : the independent sets of the matroid.

(8)

Uniform matroids

The uniform matroid U k,n is the matroid on n elements whose independent sets are all the subsets of size k.

Example 1.2

Let E = {1}. One hase two uniform matroids

U 0,1 = (E , {∅}); U 1,1 = (E , {∅, {1}}).

Example 1.3

Let E = {1, 2} and I = {∅, {1}, {2}}. One has the uniform matroid U 1,2 .

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 4 / 18

(9)

Uniform matroids

The uniform matroid U k,n is the matroid on n elements whose independent sets are all the subsets of size k.

Example 1.2

Let E = {1}. One hase two uniform matroids

U 0,1 = (E , {∅});

U 1,1 = (E , {∅, {1}}).

Example 1.3

Let E = {1, 2} and I = {∅, {1}, {2}}. One has the uniform matroid U 1,2 .

(10)

Uniform matroids

The uniform matroid U k,n is the matroid on n elements whose independent sets are all the subsets of size k.

Example 1.2

Let E = {1}. One hase two uniform matroids

U 0,1 = (E , {∅});

U 1,1 = (E , {∅, {1}}).

Example 1.3

Let E = {1, 2} and I = {∅, {1}, {2}}. One has the uniform matroid U 1,2 .

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 4 / 18

(11)

Uniform matroids

The uniform matroid U k,n is the matroid on n elements whose independent sets are all the subsets of size k.

Example 1.2

Let E = {1}. One hase two uniform matroids

U 0,1 = (E , {∅});

U 1,1 = (E , {∅, {1}}).

Example 1.3

Let E = {1, 2} and I = {∅, {1}, {2}}. One has the uniform matroid U 1,2 .

(12)

Rank function

Definition 1.4

Let M = (E , I ) be a matroid and A ⊂ E . The rank function of A:

r(A) = max{|B| : B ∈ I, B ⊂ A}. (1)

The nullity function of A:

n(A) = |A| − r (A). (2)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 5 / 18

(13)

Rank function

Definition 1.4

Let M = (E , I ) be a matroid and A ⊂ E . The rank function of A:

r(A) = max{|B| : B ∈ I, B ⊂ A}. (1)

The nullity function of A:

n(A) = |A| − r (A). (2)

(14)

Loop-coloop

Definition 1.5

Let e ∈ E . The element e is called a loop if r({e}) = 0.

The element e is called a coloop if r (E − {e}) = r (E) − 1.

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 6 / 18

(15)

Loop-coloop

Definition 1.5

Let e ∈ E . The element e is called a loop if r({e}) = 0.

The element e is called a coloop if r (E − {e}) = r (E) − 1.

(16)

Two operations [1]

Definition 1.6 (Deletion)

One sets the collection of subsets that

I 0 = {I ⊂ E − T : I ∈ I}. (3) Then one has that the pair (E − T , I 0 ) is a matroid, called that the deletion of T from M.

, → One denotes that M\ T

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 7 / 18

(17)

Two operations [2]

Definition 1.7 (Contraction)

One sets the collection of subsets that

I 00 = {I ⊂ E − T : I ∪ B T ∈ I}, (4) where B T is a maximal independent subset of T .

Then one has that the pair (E − T , I 00 ) is a matroid, called that the contraction of T from M.

, → One denotes that M/ T

(18)

Tutte polynomial for matroids

Definition 1.8

The Tutte polynomial of matroid M : T M (x , y ) = X

A⊆E

(x − 1) r(E)−r(A) (y − 1) n(A) . (5)

Example 1.9

T U

1,2

(x, y ) = x + y. (6)

Theorem 1.10

Deletion-contraction relation:

T M (x, y) = T M/e (x, y) + T M\e (x , y ). (7)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 9 / 18

(19)

Tutte polynomial for matroids

Definition 1.8

The Tutte polynomial of matroid M : T M (x , y ) = X

A⊆E

(x − 1) r(E)−r(A) (y − 1) n(A) . (5)

Example 1.9

T U

1,2

(x, y ) = x + y. (6)

Theorem 1.10

Deletion-contraction relation:

T M (x, y) = T M/e (x, y) + T M\e (x , y ). (7)

(20)

Tutte polynomial for matroids

Definition 1.8

The Tutte polynomial of matroid M : T M (x , y ) = X

A⊆E

(x − 1) r(E)−r(A) (y − 1) n(A) . (5)

Example 1.9

T U

1,2

(x, y ) = x + y. (6)

Theorem 1.10

Deletion-contraction relation:

T M (x, y) = T M/e (x, y) + T M\e (x , y ). (7)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 9 / 18

(21)

Hopf algebra on matroids

(H. Crapo and W. Schmitt. A free subalgebra of the algebra of matroids. EJC, 26(7), 05.)

Coproduct

∆(M ) = X

A⊆E

M|A ⊗ M /A. (8)

(M) =

( 1, if M = U 0,0 , 0 otherwise.

∆(U 1,2 ) = 1 ⊗ U 1,2 + 2U 1,1 ⊗ U 0,1 + U 1,2 ⊗ 1.

(k( M), f ⊕, 1, ∆, ) is bialgebra.

Moreover, this bialgebra is graded by the cardinal of the ground set, then it is a

Hopf algebra.

(22)

Two infinitesimal characters

Let us define two linear forms.

δ loop (M ) =

( 1 K if M = U 0,1 ,

0 K otherwise . (9)

δ coloop (M) =

( 1 K if M = U 1,1 ,

0 K otherwise . (10)

One has

δ loop (M 1 ⊕ M 2 ) = δ loop (M 1 )(M 2 ) + (M 1 )δ loop (M 2 ). δ coloop (M 1 ⊕ M 2 ) = δ coloop (M 1 )(M 2 ) + (M 1coloop (M 2 ).

Theorem 2.1

exp {aδ coloop + bδ loop } is a Hopf algebra character.

exp {aδ coloop + bδ loop }(M) = a r(M) b n(M) . (11)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 11 / 18

(23)

Two infinitesimal characters

Let us define two linear forms.

δ loop (M ) =

( 1 K if M = U 0,1 ,

0 K otherwise . (9)

δ coloop (M) =

( 1 K if M = U 1,1 ,

0 K otherwise . (10)

One has

δ loop (M 1 ⊕ M 2 ) = δ loop (M 1 )(M 2 ) + (M 1 )δ loop (M 2 ).

δ coloop (M 1 ⊕ M 2 ) = δ coloop (M 1 )(M 2 ) + (M 1coloop (M 2 ).

Theorem 2.1

exp {aδ coloop + bδ loop } is a Hopf algebra character.

exp {aδ coloop + bδ loop }(M) = a r(M) b n(M) . (11)

(24)

Two infinitesimal characters

Let us define two linear forms.

δ loop (M ) =

( 1 K if M = U 0,1 ,

0 K otherwise . (9)

δ coloop (M) =

( 1 K if M = U 1,1 ,

0 K otherwise . (10)

One has

δ loop (M 1 ⊕ M 2 ) = δ loop (M 1 )(M 2 ) + (M 1 )δ loop (M 2 ).

δ coloop (M 1 ⊕ M 2 ) = δ coloop (M 1 )(M 2 ) + (M 1coloop (M 2 ).

Theorem 2.1

exp {aδ coloop + bδ loop } is a Hopf algebra character.

exp {aδ coloop + bδ loop }(M) = a r(M) b n(M) . (11)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 11 / 18

(25)

A mapping α

Let us define

α(x, y, s, M ) := exp s{δ coloop + (y − 1)δ loop }

∗exp s{(x − 1)δ coloop + δ loop }(M). (12)

Proposition 3.1

α is a Hopf algebra character. Moreover, one has

α(x , y , s, M) = s |E| T M (x, y). (13)

(26)

A mapping α

Let us define

α(x, y, s, M ) := exp s{δ coloop + (y − 1)δ loop }

∗exp s{(x − 1)δ coloop + δ loop }(M). (12)

Proposition 3.1

α is a Hopf algebra character. Moreover, one has

α(x , y , s, M) = s |E| T M (x, y). (13)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 12 / 18

(27)

A convolution formula for Tutte polynomials

The character α can be rewritten:

α(x, y, s, M ) = exp (s(δ coloop + (y − 1)δ loop )) ∗ exp (s(−δ coloop + δ loop ))

∗ exp (s(δ coloop − δ loop )) ∗ exp (s((x − 1)δ coloop + δ loop )) (14) .

Corollary 3.2 (Theorem 1 of [KRS99])

The Tutte polynomial satisfies T M (x, y) = X

A⊂E

T M|A (0, y )T M/A (x , 0). (15)

W. Kook, V. Reiner, and D. Stanton. A Convolution Formula for the Tutte

Polynomial. Journal of Combinatorial Series (99)

(28)

A convolution formula for Tutte polynomials

The character α can be rewritten:

α(x, y, s, M ) = exp (s(δ coloop + (y − 1)δ loop )) ∗ exp (s(−δ coloop + δ loop ))

∗ exp (s(δ coloop − δ loop )) ∗ exp (s((x − 1)δ coloop + δ loop )) (14) .

Corollary 3.2 (Theorem 1 of [KRS99])

The Tutte polynomial satisfies T M (x, y) = X

A⊂E

T M|A (0, y )T M/A (x , 0). (15)

W. Kook, V. Reiner, and D. Stanton. A Convolution Formula for the Tutte Polynomial. Journal of Combinatorial Series (99)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 13 / 18

(29)

Differential equation of character α

Proposition 4.1

The character α is the solution of the differential equation:

ds (M ) = (xα ∗ δ coloop + y δ loop ∗ α + [δ coloop , α] − [δ loop , α] )(M ). (16)

(30)

We take a four-variable matroid polynomial Q M (x, y, a, b) which has the following properties:

a multiplicative law

Q M

1

⊕M

2

(x, y, a, b) = Q M

1

(x, y, a, b)Q M

2

(x, y, a, b), (17) if e is a coloop, then

Q M (x, y, a, b) = xQ M\e (x , y , a, b), (18) if e is a loop, then

Q M (x, y, a, b) = yQ M/e (x , y , a, b), (19) if e is a nonseparating point, then

Q M (x, y , a, b) = aQ M\e (x, y , a, b) + bQ M/e (x , y , a, b). (20)

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 15 / 18

(31)

A mapping β

β(x, y, a, b, s, M ) := s |E| Q M (x, y , a, b). (21)

Lemma 4.2

The mapping β is a matroid Hopf algebra character.

Proposition 4.3

The character β satisfies the following differential equation: dβ

ds (M) = (xβ ∗ δ coloop + yδ loop ∗ β + b[δ coloop , β] − a[δ loop , β] ) (M ). (22)

Sketch of the proof: Using the definitions of the infitesimal character δ loop and

δ coloop and the conditions of the polynomial Q M (x, y , a, b), one can get the result.

(32)

A mapping β

β(x, y, a, b, s, M ) := s |E| Q M (x, y , a, b). (21)

Lemma 4.2

The mapping β is a matroid Hopf algebra character.

Proposition 4.3

The character β satisfies the following differential equation: dβ

ds (M) = (xβ ∗ δ coloop + yδ loop ∗ β + b[δ coloop , β] − a[δ loop , β] ) (M ). (22) Sketch of the proof: Using the definitions of the infitesimal character δ loop and δ coloop and the conditions of the polynomial Q M (x, y , a, b), one can get the result.

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 16 / 18

(33)

A mapping β

β(x, y, a, b, s, M ) := s |E| Q M (x, y , a, b). (21)

Lemma 4.2

The mapping β is a matroid Hopf algebra character.

Proposition 4.3

The character β satisfies the following differential equation:

ds (M) = (xβ ∗ δ coloop + y δ loop ∗ β + b[δ coloop , β] − a[δ loop , β] ) (M ). (22)

Sketch of the proof: Using the definitions of the infitesimal character δ loop and

δ coloop and the conditions of the polynomial Q M (x, y , a, b), one can get the result.

(34)

A mapping β

β(x, y, a, b, s, M ) := s |E| Q M (x, y , a, b). (21)

Lemma 4.2

The mapping β is a matroid Hopf algebra character.

Proposition 4.3

The character β satisfies the following differential equation:

ds (M) = (xβ ∗ δ coloop + y δ loop ∗ β + b[δ coloop , β] − a[δ loop , β] ) (M ). (22) Sketch of the proof: Using the definitions of the infitesimal character δ loop and δ coloop and the conditions of the polynomial Q M (x, y , a, b), one can get the result.

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 16 / 18

(35)

Main theorem

From the propositions 3.1, 4.1 and 4.3, one gets the result

Theorem 4.4

Q(x, y, a, b, M ) = a n(M) b r(M) T M ( x b , y

a ). (23)

(36)

Thank you for your attention!

Nguyen Hoang-Nghia (LIPN, Universit´e Paris 13) Recipe theorem for the Tutte polynomial for matroids, renormalization group-like approachEllwangen, SLC 70 18 / 18

参照

関連したドキュメント

Following Speyer, we give a non-recursive formula for the bounded octahedron recurrence using perfect matchings.. Namely, we prove that the solution of the recur- rence at some

As is well known (see [20, Corollary 3.4 and Section 4.2] for a geometric proof), the B¨ acklund transformation of the sine-Gordon equation, applied repeatedly, produces

[18] , On nontrivial solutions of some homogeneous boundary value problems for the multidi- mensional hyperbolic Euler-Poisson-Darboux equation in an unbounded domain,

Since the boundary integral equation is Fredholm, the solvability theorem follows from the uniqueness theorem, which is ensured for the Neumann problem in the case of the

Next we integrate out all original domain wall indices α, β, γ, · · · and we get the effective weight function of M at each coarse grained (renormalized) dual link, where M is

In 1965, Kolakoski [7] introduced an example of a self-generating sequence by creating the sequence defined in the following way..

Next, we prove bounds for the dimensions of p-adic MLV-spaces in Section 3, assuming results in Section 4, and make a conjecture about a special element in the motivic Galois group

Transirico, “Second order elliptic equations in weighted Sobolev spaces on unbounded domains,” Rendiconti della Accademia Nazionale delle Scienze detta dei XL.. Memorie di