• 検索結果がありません。

A survey on Stanley decompositions

N/A
N/A
Protected

Academic year: 2022

シェア "A survey on Stanley decompositions"

Copied!
59
0
0

読み込み中.... (全文を見る)

全文

(1)

A survey on Stanley decompositions

J ¨urgen Herzog Universit ¨at Duisburg-Essen

Ellwangen, March 2011

(2)

Outline

The conjecture

Known cases

How to compute the Stanley depth

Upper and lower bounds

Stanley depth for syzygies

(3)

Outline

The conjecture

Known cases

How to compute the Stanley depth

Upper and lower bounds

Stanley depth for syzygies

(4)

Outline

The conjecture

Known cases

How to compute the Stanley depth

Upper and lower bounds

Stanley depth for syzygies

(5)

Outline

The conjecture

Known cases

How to compute the Stanley depth

Upper and lower bounds

Stanley depth for syzygies

(6)

Outline

The conjecture

Known cases

How to compute the Stanley depth

Upper and lower bounds

Stanley depth for syzygies

(7)

The conjecture

Richard Stanley in his article “Linear Diophantine equations and local cohomology”, Invent. Math. 68 (1982) made a striking conjecture concerning the depth of multigraded modules.

(8)

The conjecture

Richard Stanley in his article “Linear Diophantine equations and local cohomology”, Invent. Math. 68 (1982) made a striking conjecture concerning the depth of multigraded modules.

Here we concentrate on the case that M is a finitely generated Zn-graded S-module, whereS =K[x1, . . . ,xn]is the polynomial ring.

(9)

The conjecture

Richard Stanley in his article “Linear Diophantine equations and local cohomology”, Invent. Math. 68 (1982) made a striking conjecture concerning the depth of multigraded modules.

Here we concentrate on the case that M is a finitely generated Zn-graded S-module, whereS =K[x1, . . . ,xn]is the polynomial ring.

An important special case for aZn-graded S-module is M=I/J where JIS are monomial ideals.

(10)

AStanley decompositionDof M is direct sum ofZn-graded K -vector spaces

D:M=

r

M

j=1

mjK[Zj],

where each mjM is homogeneous, ZjX ={x1, . . . ,xn} and each mjK[Zj]is a free K[Zj]-module.

(11)

AStanley decompositionDof M is direct sum ofZn-graded K -vector spaces

D:M=

r

M

j=1

mjK[Zj],

where each mjM is homogeneous, ZjX ={x1, . . . ,xn} and each mjK[Zj]is a free K[Zj]-module.

We set sdepth(D) =min{|Zj| j =1, . . . ,r}, and

sdepth M =max{sdepth(D) :Dis a Stanley decomposition of M}. is called theStanley depthof M.

(12)

AStanley decompositionDof M is direct sum ofZn-graded K -vector spaces

D:M=

r

M

j=1

mjK[Zj],

where each mjM is homogeneous, ZjX ={x1, . . . ,xn} and each mjK[Zj]is a free K[Zj]-module.

We set sdepth(D) =min{|Zj| j =1, . . . ,r}, and

sdepth M =max{sdepth(D) :Dis a Stanley decomposition of M}. is called theStanley depthof M.

Conjecture (Stanley)sdepth Mdepth M.

(13)

Example: I = (x1x23,x13x2)

• •

The figure displays Stanley decompositions of

I=x1x23K[x1,x2]⊕x13x22K[x1]⊕x13x2K[x1], and

S/I=K[x2]⊕x1K[x1]⊕x1x2Kx1x22Kx12x2Kx12x22K.

(14)

Known cases

The Stanley depth for modules of the form I/J where JISK[x1, . . . ,xn]are monomial ideals is apure

combinatorial invariant, in particular, it does not depend on the field K , while the depth ishomological invariantand in case of squarefree monomial ideal, a topological invariant of the attached simplicial complex, and may very well depend on the field K .

(15)

Known cases

The Stanley depth for modules of the form I/J where JISK[x1, . . . ,xn]are monomial ideals is apure

combinatorial invariant, in particular, it does not depend on the field K , while the depth ishomological invariantand in case of squarefree monomial ideal, a topological invariant of the attached simplicial complex, and may very well depend on the field K .

What is know for I/J?

(16)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(17)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(H, Popescu) If∆is a simplicial complex. Then the Stanley Reisner ringK[∆]of∆satisfies Stanley’s conjecture, if∆ is shellable.

(18)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(H, Popescu) If∆is a simplicial complex. Then the Stanley Reisner ringK[∆]of∆satisfies Stanley’s conjecture, if∆ is shellable.

(H, Jahan, Yassemi) Stanley’s conjecture holds forS/I whenIis Cohen-Macaulay ideal of codimension 2 or Gorenstein of codimension 3.

(19)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(H, Popescu) If∆is a simplicial complex. Then the Stanley Reisner ringK[∆]of∆satisfies Stanley’s conjecture, if∆ is shellable.

(H, Jahan, Yassemi) Stanley’s conjecture holds forS/I whenIis Cohen-Macaulay ideal of codimension 2 or Gorenstein of codimension 3.

(Popescu) IfSis a polynomial ring in at most 5 variables, then Stanley’s conjecture holds forS/I.

(20)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(H, Popescu) If∆is a simplicial complex. Then the Stanley Reisner ringK[∆]of∆satisfies Stanley’s conjecture, if∆ is shellable.

(H, Jahan, Yassemi) Stanley’s conjecture holds forS/I whenIis Cohen-Macaulay ideal of codimension 2 or Gorenstein of codimension 3.

(Popescu) IfSis a polynomial ring in at most 5 variables, then Stanley’s conjecture holds forS/I.

(Cimpoeas) IfIS=K[x1, . . . ,xn]is generated by at most 2n−1monomials, then Stanley’s conjecture holds forI

(21)

(Jahan, Zheng, H) Stanley’s conjecture holds for all algebrasS/I,Ia monomial ideal, if it holds for all such Cohen–Macaulay algebras.

(H, Popescu) If∆is a simplicial complex. Then the Stanley Reisner ringK[∆]of∆satisfies Stanley’s conjecture, if∆ is shellable.

(H, Jahan, Yassemi) Stanley’s conjecture holds forS/I whenIis Cohen-Macaulay ideal of codimension 2 or Gorenstein of codimension 3.

(Popescu) IfSis a polynomial ring in at most 5 variables, then Stanley’s conjecture holds forS/I.

(Cimpoeas) IfIS=K[x1, . . . ,xn]is generated by at most 2n−1monomials, then Stanley’s conjecture holds forI

(Apel, Okazaki, Yanagawa) IfIis a cogeneric monomial ideal, then Stanley’s conjecture holds forS/I.

(22)

How to compute the Stanley depth

We return to the case M =I/J where JIS=K[x1, . . . ,xn] are monomial ideals.

(23)

How to compute the Stanley depth

We return to the case M =I/J where JIS=K[x1, . . . ,xn] are monomial ideals.

We choose g ∈Nnsuch that ga for all generators xaof I and J, and consider the finite poset

PI/Jg ={a∈Nn xaI\J, ag}.

We call it thecharacteristic posetof I/J with respect to g.

(24)

How to compute the Stanley depth

We return to the case M =I/J where JIS=K[x1, . . . ,xn] are monomial ideals.

We choose g ∈Nnsuch that ga for all generators xaof I and J, and consider the finite poset

PI/Jg ={a∈Nn xaI\J, ag}.

We call it thecharacteristic posetof I/J with respect to g.

• • •

••

••

••

• •

• •

(25)

How to compute the Stanley depth

We return to the case M =I/J where JIS=K[x1, . . . ,xn] are monomial ideals.

We choose g ∈Nnsuch that ga for all generators xaof I and J, and consider the finite poset

PI/Jg ={a∈Nn xaI\J, ag}.

We call it thecharacteristic posetof I/J with respect to g.

• • •

••

••

••

• •

• •

If∆is a simplicial complex andg = (1, . . . ,1), thenPK[∆]g can be identified with the face poset of∆.

(26)

The characteristic poset ofm= (x1,x2,x3)with respect to g = (1,1,1)is given by

◦ ◦ ◦

◦ ◦ ◦

1 2 3

12 23 13

123

(27)

Given any poset P and a,bP. Then

[a,b] ={cP : acb}is called aninterval.

(28)

Given any poset P and a,bP. Then

[a,b] ={cP : acb}is called aninterval.

Suppose P is a finite poset. Apartitionof P is a disjoint union

P : P=

r

[

i=1

[ai,bi]

of intervals.

(29)

Given any poset P and a,bP. Then

[a,b] ={cP : acb}is called aninterval.

Suppose P is a finite poset. Apartitionof P is a disjoint union

P : P=

r

[

i=1

[ai,bi]

of intervals.

P : Pmg = [1,12]∪[2,23]∪[3,13]∪[123,123].

is a partition of Pmg.

(30)

◦ ◦ ◦

◦ ◦ ◦

1 2 3

12 23 13

123

(31)

Each partition of PI/Jg gives rise to a Stanley decomposition of I/J.

(32)

Each partition of PI/Jg gives rise to a Stanley decomposition of I/J.

In order to describe the Stanley decomposition of I/J coming from a partition of PI/Jg we shall need the following notation: for each bPI/Jg , we set

Zb={xj : b(j) =g(j)},

(33)

Each partition of PI/Jg gives rise to a Stanley decomposition of I/J.

In order to describe the Stanley decomposition of I/J coming from a partition of PI/Jg we shall need the following notation: for each bPI/Jg , we set

Zb={xj : b(j) =g(j)}, and define

ρ PI/Jg →Z≥0, b7→ρ(b) =|Zb|.

(34)

Theorem (a) LetP : PI/Jg =Sr

i=1[ci,di]be a partition ofPI/Jg . Then

D(P) : I/J =

r

M

i=1

(M

c

xcK[Zdi])

is a Stanley decomposition of I/J, where the inner direct sum is taken over allc∈[ci,di]for which c(j) =ci(j)for all j with xjZdi. Moreover,sdepthD(P) =min{ρ(di) i=1, . . . ,r}.

(35)

Theorem (a) LetP : PI/Jg =Sr

i=1[ci,di]be a partition ofPI/Jg . Then

D(P) : I/J =

r

M

i=1

(M

c

xcK[Zdi])

is a Stanley decomposition of I/J, where the inner direct sum is taken over allc∈[ci,di]for which c(j) =ci(j)for all j with xjZdi. Moreover,sdepthD(P) =min{ρ(di) i=1, . . . ,r}. (b) LetDbe a Stanley decomposition of I/J. Then there exists a partitionP of PI/Jg such that

sdepthD(P)≥sdepthD.

In particular, sdepth I/J can be computed as the maximum of the numbers sdepthD(P), whereP runs over the (finitely many) partitions of PI/Jg .

(36)

• •

• • •

• • •

••

••

••

• •

• •

••

••

• •

(37)

This theorem has been used to compute or to estimate the Stanley depth in several cases:

(C. Biro, D. Howard, M. Keller, W. Trotter, S. Young) sdepth(x1, . . . ,xn) =⌈n/2⌉.

(38)

This theorem has been used to compute or to estimate the Stanley depth in several cases:

(C. Biro, D. Howard, M. Keller, W. Trotter, S. Young) sdepth(x1, . . . ,xn) =⌈n/2⌉.

(Shen) LetIS=K[x1, . . . ,xn]be a complete intersection monomial ideal minimally generated by m elements. Thensdepth(I) =n− ⌊m/2⌋.

(39)

This theorem has been used to compute or to estimate the Stanley depth in several cases:

(C. Biro, D. Howard, M. Keller, W. Trotter, S. Young) sdepth(x1, . . . ,xn) =⌈n/2⌉.

(Shen) LetIS=K[x1, . . . ,xn]be a complete intersection monomial ideal minimally generated by m elements. Thensdepth(I) =n− ⌊m/2⌋.

(Floystad, H) Letsbe the largest integer such that n+1≥(2s+1)(s+1).Then the Stanley depth of any squarefree monomial ideal in n variables is greater or equal to 2s+1. Explicitly this lower bound is

2

2n+2.25+0.5 2

−1.

(40)

Upper and lower bounds

Let M be aZn-graded S=K[x1, . . . ,xn]-module. Then there exists

F : 0=M0M1⊂ · · · ⊂Mm=M a chain ofZn-graded submodules of M such that

Mi/Mi−1≃(S/Pi)(−ai)where ai ∈Znand where each Pi is a monomial prime ideal.

(41)

Upper and lower bounds

Let M be aZn-graded S=K[x1, . . . ,xn]-module. Then there exists

F : 0=M0M1⊂ · · · ⊂Mm=M a chain ofZn-graded submodules of M such that

Mi/Mi−1≃(S/Pi)(−ai)where ai ∈Znand where each Pi is a monomial prime ideal.

One has

Min(M)⊂Ass(M)⊂ {P1, . . . ,Pr} ⊂Supp(M), and

(42)

Upper and lower bounds

Let M be aZn-graded S=K[x1, . . . ,xn]-module. Then there exists

F : 0=M0M1⊂ · · · ⊂Mm=M a chain ofZn-graded submodules of M such that

Mi/Mi−1≃(S/Pi)(−ai)where ai ∈Znand where each Pi is a monomial prime ideal.

One has

Min(M)⊂Ass(M)⊂ {P1, . . . ,Pr} ⊂Supp(M), and

min{dim S/P1, . . . ,S/Pr} ≤ depth M,sdepth M

≤ min{dim S/P: P ∈Ass(M)}. The upper inequality has been proved by Apel.

(43)

Let M be a finitely generatedZn-graded

S =K[x1, . . . ,xn]-module. It is alsoZ-graded,i.e.,

M =M

i∈Z

Mi, with Mi = M

a∈Zn

|a|=i

Ma.

(44)

Let M be a finitely generatedZn-graded

S =K[x1, . . . ,xn]-module. It is alsoZ-graded,i.e.,

M =M

i∈Z

Mi, with Mi = M

a∈Zn

|a|=i

Ma.

HM(t) =X

i∈Z

dimKMiti

(45)

Let M be a finitely generatedZn-graded

S =K[x1, . . . ,xn]-module. It is alsoZ-graded,i.e.,

M =M

i∈Z

Mi, with Mi = M

a∈Zn

|a|=i

Ma.

HM(t) =X

i∈Z

dimKMiti = Q(t)

(1−t)d, Q(1)6=0 is the Hilbert series of M.

(46)

Let M be a finitely generatedZn-graded

S =K[x1, . . . ,xn]-module. It is alsoZ-graded,i.e.,

M =M

i∈Z

Mi, with Mi = M

a∈Zn

|a|=i

Ma.

HM(t) =X

i∈Z

dimKMiti = Q(t)

(1−t)d, Q(1)6=0 is the Hilbert series of M.

Example: HS(t) = (1−t1)n forS=K[x1, . . . ,xn].

(47)

Given a Stanley decomposition

D: M=

r

M

j=1

mjK[Zj], deg mj =aj, bj =|Zj|.

(48)

Given a Stanley decomposition

D: M=

r

M

j=1

mjK[Zj], deg mj =aj, bj =|Zj|.

One obtains

S : HM(t) =

r

X

j=1

taj (1−t)bj.

(49)

Given a Stanley decomposition

D: M=

r

M

j=1

mjK[Zj], deg mj =aj, bj =|Zj|.

One obtains

S : HM(t) =

r

X

j=1

taj (1−t)bj.

For any such sum decompositionS of HM(t)we set

hdepthS =min{b1, . . . ,rr}

, and calls

hdepth M =max{hdepthS : S is a sum decomposition of HM(t)}. theHilbert depthof M.

(50)

Given a Stanley decomposition

D: M=

r

M

j=1

mjK[Zj], deg mj =aj, bj =|Zj|.

One obtains

S : HM(t) =

r

X

j=1

taj (1−t)bj.

For any such sum decompositionS of HM(t)we set

hdepthS =min{b1, . . . ,rr}

, and calls

hdepth M =max{hdepthS : S is a sum decomposition of HM(t)}. theHilbert depthof M. Obviously,hdepth Msdepth M.

(51)

The inequalityhdepth Mdepth M is unknown in general.

(52)

The inequalityhdepth Mdepth M is unknown in general.

The Hilbert depth has been computed is some interesting special cases

(Bruns, Krattenthaler, Uliczka) LetM(n,k)be the k -syzygy module ofK =S/(x1, . . . ,xn). Then

hdepth M(n,k) =n−1 forn/2⌋ ≤k <n,

and

hdepth M(n,k)≥sdepth M(n,k) ≥ ⌊(n+k)/2⌋ for k <n/2.

(53)

The inequalityhdepth Mdepth M is unknown in general.

The Hilbert depth has been computed is some interesting special cases

(Bruns, Krattenthaler, Uliczka) LetM(n,k)be the k -syzygy module ofK =S/(x1, . . . ,xn). Then

hdepth M(n,k) =n−1 forn/2⌋ ≤k <n,

and

hdepth M(n,k)≥sdepth M(n,k) ≥ ⌊(n+k)/2⌋

for k <n/2.

(Bruns, Krattenthaler, Ulizcka) hdepth(x1, . . . ,xn)k =⌈n/(k +1)⌉

(54)

Stanley depth for syzygies

We have

hdepth M(n,k) =n−1≥depth M(n,k) =k for⌊n/2⌋ ≤k <n.

(55)

Stanley depth for syzygies

We have

hdepth M(n,k) =n−1≥depth M(n,k) =k

for⌊n/2⌋ ≤k <n.

In general, if M is a finitely generatedZn-graded of depth t, and Zk(M)is its k th syzygy module, then

depth Zk(M) =t+k for k =1, . . . ,nt.

In particular for IS one hasdepth Idepth S/I+1.

(56)

Stanley depth for syzygies

We have

hdepth M(n,k) =n−1≥depth M(n,k) =k

for⌊n/2⌋ ≤k <n.

In general, if M is a finitely generatedZn-graded of depth t, and Zk(M)is its k th syzygy module, then

depth Zk(M) =t+k for k =1, . . . ,nt.

In particular for IS one hasdepth Idepth S/I+1.

It is not known whethersdepth I =sdepth S/I+1.

(57)

Stanley depth for syzygies

We have

hdepth M(n,k) =n−1≥depth M(n,k) =k

for⌊n/2⌋ ≤k <n.

In general, if M is a finitely generatedZn-graded of depth t, and Zk(M)is its k th syzygy module, then

depth Zk(M) =t+k for k =1, . . . ,nt.

In particular for IS one hasdepth Idepth S/I+1.

It is not known whethersdepth I =sdepth S/I+1.

Theorem (Floystad, H)sdepth Zk(M)≥k for all k .

(58)

J. Apel, On a conjecture of R. P. Stanley; Part I - Monomial Ideals, J. of Alg. Comb. 17, (2003), 3956.

J. Apel, On a conjecture of R. P. Stanley; Part II - Quotients Modulo Monomial Ideals, J. of Alg. Comb. 17, (2003), 5774.

W.Bruns, J.Herzog, Cohen–Macaualy rings, Cambridge University Press, Revised edition (1998)

W.Bruns, C.Krattenthaler, J.Ulizka, Stanley depth and Hilbert decompositions in the Koszul complex,

arXiv:0909.0686.

W. Bruns, C.Krattenthaler, J.Uliczka, Hilbert depth of powers of the maximal ideal, arXiv:1002.1400

M. Cimpoeas, A note on Stanleys conjecture for monomial ideals, arXiv:0906.1303

M.Keller, Y.-H.Shen, N.Streib, S.Young, On the Stanley depth of squarefree Veronese ideals, arXiv:0910.4645.

G.Floystad, J. Herzog, Gr ¨obner bases of syzygies and Stanley depth.

(59)

J.Herzog, M.Vladiou, X.Zheng, How to compute the Stanley depth of a monomial ideal, J. Algebra 322 (2009),

3151–3169.

J. Herzog, A. SoleymanJahan and S. Yassemi, Stanley decompositions and partitionable simplicial complexes. J.

Algebr. Comb. 27, 113125 (2008).

J. Herzog, A. Soleyman Jahan, X. Zheng, Skeletons of monomial ideals, Math. Nachr. 283 (2010), 14031408.

R. Okazaki, K. Yanagawa, Alexander duality and Stanley depth of multigraded modules.

D. Popescu, Stanley depth of multigraded modules, J.

Algebra 321(2009), 2782-2797.

Y. Shen, Stanley depth of complete intersection monomial ideals and upper-discrete partitions, J. Algebra 321 (2009), 12851292.

R.Stanley, Linear Diophantine equations and local cohomology, Inventiones Mathematicae 68, (1982), 175–193.

参照

関連したドキュメント

[19, 20], and it seems to be commonly adopted now.The general background for these geometries goes back to Klein’s definition of geometry as the study of homogeneous spaces, which

In my earlier paper [H07] and in my talk at the workshop on “Arithmetic Algebraic Geometry” at RIMS in September 2006, we made explicit a conjec- tural formula of the L -invariant

In this paper a similar problem is studied for semidynamical systems. We prove that a non-trivial, weakly minimal and negatively strongly invariant sets in a semidynamical system on

Polynomial invariant and reciprocity theorem on the Hopf monoid of hypergraphs..

This applies to the case where the induced action 1 ϕ acts transitively on the base manifold and states that each point in the bundle gives rise to a bijection between the set

We study the classical invariant theory of the B´ ezoutiant R(A, B) of a pair of binary forms A, B.. We also describe a ‘generic reduc- tion formula’ which recovers B from R(A, B)

Key words and phrases: Quasianalytic ultradistributions; Convolution of ultradistributions; Translation-invariant Banach space of ultradistribu- tions; Tempered

In secgion 3 we prove a general theorem which ensures ghe exisgence and uniqueness of invarian measures for McKean-Vlasov nonlinear stochastic differential equations.. In section 4