• Aucun résultat trouvé

(1)MIRCEA CIMPOEAS¸ Communicated by Vasile Brˆınz˘anescu This is a survey of the results obtained by the author regarding the Stanley depth of monomial ideals and the Stanley’s conjecture

N/A
N/A
Protected

Academic year: 2022

Partager "(1)MIRCEA CIMPOEAS¸ Communicated by Vasile Brˆınz˘anescu This is a survey of the results obtained by the author regarding the Stanley depth of monomial ideals and the Stanley’s conjecture"

Copied!
8
0
0

Texte intégral

(1)

MIRCEA CIMPOEAS¸ Communicated by Vasile Brˆınz˘anescu

This is a survey of the results obtained by the author regarding the Stanley depth of monomial ideals and the Stanley’s conjecture.

AMS 2010 Subject Classification: Primary: 13H10, Secondary: 13P10.

Key words: Stanley depth, Stanley’s conjecture, monomial ideal.

1. INTRODUCTION

Let K be a field and S=K[x1, . . . , xn] the polynomial ring overK. Let M be a Zn-gradedS-module. A Stanley decomposition of M is a direct sum D:M =Lr

i=1miK[Zi] asK-vector space, where mi ∈M,Zi ⊂ {x1, . . . , xn} such that miK[Zi] is a free K[Zi]-module. We define sdepth(D) = minri=1|Zi| and sdepthS(M) = max{sdepth(D)| D is a Stanley decomposition of M}, see [1]. The number sdepth(M) is called theStanley depth ofM. Herzog, Vladoiu and Zheng show in [11] that this invariant can be computed in a finite number of steps if M = I/J, where J ⊂ I ⊂ S are monomial ideals. There are two important particular cases. If I ⊂ S is a monomial ideal, we are interested in computing sdepthS(S/I) and sdepthS(I) and to find some relation between them.

Csaba Biro, David M. Howard, Mitchel T. Keller, William T. Trotter and Stephen J. Young proved that ifm= (x1, . . . , xn)⊂S, then sdepth(m) =n

2

, see [2, Theorem 2.2]. We extended this result to monomial ideals generated by powers of variables, see [7, Theorem 1.3]. More precisely, we proved in [5]

that sdepth(mk)≤l

n k+1

m

, for any positive integer k and we conjectured that the equality holds. Shen proved that if I is a complete intersection monomial ideal minimally generated by m monomials, then sdepth(I) = n−m

2

, see [17, Theorem 2.4]. Okazaki proved that if I is an arbitrary monomial ideal minimally generated by m monomials, then sdepth(I) ≥ n−m

2

, see [13, Theorem 2.1].

In [6], we compute the Stanley depth for the quotient ring of a square free Veronese ideal and we give some bounds for the Stanley depth of a square

REV. ROUMAINE MATH. PURES APPL.,58(2013),2,205–212

(2)

free Veronese ideal. In particular, it follows that both satisfy the Stanley’s conjecture. In [9] we proved that if I is a monomial ideal, almost complete intersection, then the Stanley conjecture holds for I and S/I. In [10], we give several bounds for sdepthS(I+J), sdepthS(I ∩J), sdepthS(S/(I+J)), sdepthS(S/(I∩J)), sdepthS(I :J) and sdepthS(S/(I :J)), where I, J ⊂S= K[x1, . . . , xn] are monomial ideals. Also, we give some equivalent forms of Stanley conjecture for I and S/I, whereI ⊂S is a monomial ideal.

2. STANLEY DEPTH

FOR MONOMIAL COMPLETE INTERSECTION IDEALS

We proved in [7] the following key result.

Lemma 2.1 ([8, Lemma 1.1]). Let v1, . . . , vm ∈ K[x2, . . . , xm] be some monomials and let a be a positive integer. Let I = (xa1v1, v2,· · · , vm) and I0= (xa+11 v1, v2,· · ·, vm). Then sdepth(I) = sdepth(I0).

Csaba Biro, David M. Howard, Mitchel T. Keller, William T. Trotter and Stephen J. Young proved the following theorem.

Theorem 2.2 ([2, Theorem 2.2]). If m = (x1, . . . , xn) ⊂ S, then sdepth(m) =n

2

.

As a direct consequence of Lemma 2.1 and Theorem 2.2, we get.

Theorem2.3 ([8, Theorem 1.3]). Leta1, . . . , anbe some positive integers.

Then:

sdepth((xa11, . . . , xann)) =sdepth((x1, . . . , xn)) =ln 2 m

. In particular, sdepth((xa11, . . . , xamm)) =n−m

2

for any1≤m≤n.

We also proved in [8] the following results.

Theorem 2.4 ([8, Theorem 2.1]). Let I ⊂S be a complete intersection monomial ideal. Then sdepth(I) = sdepth(√

I).

Corollary 2.5 ([8, Corrolary 1.3]). Let I ⊂S be a monomial ideal with G(I) ={v1, . . . , vm}. Let S0 =S[xn+1]and I0 = (v1, . . . , vm−1, xn+1vm).Then sdepth(I0)≤sdepth(I) + 1.

In the condition of the above Corollary, Shen proved in [17] the other inequality and, as a direct consequence he obtained.

Theorem2.6 ([17, Theorem 2.4]). LetI be a complete intersection mono- mial ideal minimally generated by m monomials. Then, sdepth(I) =n−m

2

. Okazaki gived in [13] an interesting result related with the previous.

(3)

Theorem 2.7 ([13, Theorem 2.1]). Let I ⊂S be a monomial ideal (min- imally) generated by m monomials. Then

sdepth(I)≥max{1, n−jm 2

k }.

We recall also the following result of Asia Rauf, see [16, Theorem 1.1].

Theorem2.8. LetI ⊂Sbe a monomial complete intersection, minimally generated by m monomials. Then sdepth(S/I) =n−m.

Two main results from [7] are the following.

Proposition 2.9 ([7, Proposition 1.2]). Let I ⊂S be a monomial ideal, minimally generated by m monomials. Then sdepth(S/I)≥n−m.

Note that the above Proposition is a counterpart for Theorem 2.6, since the equality sdepth(S/I) =n−m holds for a monomial complete intersection ideal I ⊂ S, minimally generated by m monomials. Therefore, monomial complete intersection ideals have the minimal Stanley depth, w.r.t. number of generators.

Theorem 2.10 ([7, Theorem 1.4]). Let I ⊂S be a monomial ideal such that I = v(I :v), for a monomial v ∈ S. Then sdepth(S/I) = sdepth(S/(I : v))and sdepth(I) = sdepth(I :v).

Theorem 2.11 ([7, Theorem 2.6]). Let I ⊂S be a monomial ideal min- imally generated by at most 3 generators. Then I and S/I satisfy the Stanley conjecture.

Our main results from [9] are the following.

Theorem 2.12 ([9, Theorem 1.8]). Let I ⊂ S = K[x1, . . . , xn] be a monomial ideal, minimally generated by m monomials, k = max{|P| : P ∈ Ass(S/I)}, and s≥k be an integer. Then

1. If m≤s−1 +l

s k−1

m

, then sdepth(S/I)≥n−s.

2. If m≤2s−3 +l

2s−2 k−1

m

, then sdepth(I)≥n−s+ 1.

If depth(S/I) =n−s then(1) and (2) imply the Stanley conjecture for S/I, respectively forI.

Corollary 2.13 ([9, Corollary 1.9]). Let I be a monomial almost com- plete intersection ideal. Then Stanley’s conjecture holds for S/I and I.

(4)

3. STANLEY DEPTH OF SQUAREFREE VERONESE IDEALS In this section we will follow [6].

Theorem 3.1 ([6, Theorem 1.1]). (1) sdepth(S/In,d) =d−1.

(2) d≤sdepth(In,d)≤ n−dd+1 +d.

Since depth(S/In,d) =d−1, we get the following corollary.

Corollary 3.2 ([6, Corollary 1.2]). In,d and S/In,d satisfy the Stanley conjecture. Also,

sdepth(In,d)≥sdepth(S/In,d) + 1.

Letk≤nbe two positive integers. We denoteAn,k ={F ⊂[n]| |F|=k}.

With this notations, we have the following well known result from combina- torics.

Theorem 3.3. For any positive integers d≤n such that d≤n/2, there exists a bijective map Φn,d :An,d →An,d, such that Φn,d(F)∩F =∅, for any F ∈An,d.

As a consequence, we get the following corollary.

Corollary 3.4 ([6, Corollary 1.5]). Let n, dbe two positive integers such that 2d+ 1≤n≤3d. Then sdepth(In,d) =d+ 1.

We also, give the following conjecture.

Conjecture 3.5. For any positive integers d ≤ n such that d ≤ n/2, sdepth(In,d) =j

n−d d+1

k +d.

In [12], Mitchel T. Keller, Yi-Huang Shen, Noah Streib and Stephen J.

Young give a patial positive answer to the above conjecture. More precisely, they proved the following result.

Theorem 3.6 ([12, Theorem 1.1]). Let In,d be the squarefree Veronese ideal in S, generated by all squarefree monomials of degree d.

(1)If1≤d≤n <5d+4, then the Stanley depthsdepth(In,d) = jn−d

d+1

k +d.

(2) If d≥1 and n≥5d+ 4 thend+ 3≤sdepth(In,d)≤j

n−d d+1

k +d.

In [5] we proved the following results, related with those above.

Theorem 3.7 ([5, Theorem 2.2]). Let k be a positive integer. Then sdepth(mk)≤

n k+ 1

. In particular, if k≥n−1, then sdepth(mk) = 1.

(5)

Proposition 3.8 ([5, Proposition 2.3]). Let I ⊂S be a monomial ideal.

Then

sdepth(mkI) = 1 f or k0.

We also conjectured that:

Conjecture 3.9. sdepth(mk) =l

n k+1

m .

Regarding this conjecture, Bruns, Krattenthaler and Uliczka proved in [4] that the Hilbert depth of the k-th power of m is exactly l

n k+1

m

. Since the Hilbert depth is always ≥than the Stanley depth we cannot conclude the equality for the Stanley depth, although the conjecture seems to be true. For the definition and basic properties of the Hilbert depth of a graded, respectively multigraded, module, see [3].

4.SEVERAL INEQUALITIES REGARDING SDEPTH In this section, we present the main result from [10]. We denote S = K[x1, . . . , xn] the ring of polynomials in n variables, where n ≥ 2. For a monomial u ∈ S, we denote supp(u) = {xi : xi|u}. We begin this section by recalling the following results. Let I ⊂ S0 = K[x1, . . . , xr], J ⊂ S00 = K[xr+1, . . . , xn] be monomial ideals, where 1 ≤ r < n. Then, we have the following inequalities:

Proposition 4.1 ([10, Proposition 1.1]).

(1) sdepthS(IS∩J S)≥sdepthS0(I) + sdepthS00(J). ([14, Lemma 1.1]).

(2) sdepthS(S/(IS +J S)) ≥ sdepthS0(S0/I) + sdepthS00(S00/J). ([15, Theorem 3.1]).

(3) depthS(S/(IS∩J S))−1 = depthS(S/(IS+J S)) = depthS0(S0/I) + depthS00(S00/J). ([14, Lemma 1.1]).

We give similar results, in the following theorem.

Theorem 4.2 ([10, Theorem 1.3]). Let I ⊂ S0 = K[x1, . . . , xr], J ⊂ S00=K[xr+1, . . . , xn]be monomial ideals, where1≤r < n. Then, we have the following inequalities:

(1) sdepthS(IS)≥sdepthS(IS+J S)≥min{sdepthS(IS),sdepthS00(J) + sdepthS0(S0/I)}.

(2) sdepthS(S/IS) ≥ sdepthS(S/(IS ∩ J S)) ≥ min{sdepthS(S/IS), sdepthS00(S00/J) + sdepthS0(I)}.

Corollary 4.3 ([5, Corrolary 1.8]). With the notations above, we have the followings:

(6)

(1)If the Stanley conjecture hold for I andJ, then the Stanley conjecture holds for IS∩J S.

(2) If the Stanley conjecture hold for S0/I and S00/J, then the Stanley conjecture holds for S/(IS+J S).

(3) If the Stanley conjecture hold for I, J andS0/I or forI, J and S00/J, then the Stanley conjecture holds for (IS+J S).

(4) If the Stanley conjecture hold forS0/I, S00/J andI orS0/I, S00/J and I and J, then the Stanley conjecture holds for S/(IS∩J S).

In the following, we consider 1≤s≤r+ 1≤nthree integers, withn≥2.

We denote S0 := K[x1, . . . , xr], S00 := K[xs, . . . , xn] and S := K[x1, . . . , xn].

Letp:=r−s+ 1. With these notations, we generalize some results of Propo- sition 4.1 and Theorem 4.2.

Theorem 4.4 ([10, Theorem 2.2]). Let I ⊂S0 and J ⊂S00 be two mono- mial ideals. Then:

(1) sdepthS(IS∩J S)≥sdepthS0(I) + sdepthS00(J)−p= sdepthS(IS) + sdepthS(J S)−n.

(2) sdepthS(S/(IS +J S)) ≥ sdepthS0(S0/I) + sdepthS00(S00/J) −p = sdepthS(S/IS) + sdepthS(S/J S)−n.

(3) sdepthS(IS+J S)≥min{sdepthS(IS),sdepthS00(J)+sdepthS0(S0/I)−

p}= min{sdepthS(IS),sdepthS(J S) + sdepthS(S/IS)−n}.

(4) sdepthS(S/(IS ∩ J S)) ≥ min{sdepthS(S/IS),sdepthS00(S00/J) + sdepthS0(I)−p}= min{sdepthS(S/IS),sdepthS(S/J S) + sdepthS(IS)−n}.

This Theorem yields different corollaries and new proofs for other results, see [10]. Now, let I ⊂ S be a monomial ideal and let I = C1 ∩ · · · ∩Ck, be the irredundant minimal decomposition of I. If we denote Pj = p

Cj for 1≤j ≤k, we have Ass(S/I) ={P1, . . . , Pk}. In particular, if I is squarefree, Cj = Pj for all j. Denote dj = ht(Pj), where 1 ≤ i ≤ k. We may assume that d1 ≥d2 ≥ · · · ≥ dk. We obtain the following bounds for sdepthS(I) and sdepthS(S/I).

Corollary 4.5 ([10, Corollary 2.13]).

(1) n− bd1/2c ≥sdepthS(I)≥n− bd1/2c − · · · − bdk/2c.

(2) n−d1≥sdepthS(S/I)≥n− bd1/2c − · · · − bdk−1/2c −dk.

In a more general case, letI =Q1∩ · · · ∩Qk be the primary irredundant decomposition of I,Pi =√

Qi and denote qj = sdepthS(Qj) and dj =ht(Pj).

We may assume that d1 ≥ d2 ≥ · · · ≥ dk. Note that qj ≤ n−dj/2, since Pj = (Qj : uj), where uj ∈ S is a monomial, and therefore sdepthS(Qj) ≤ sdepthS(Pj), by Proposition 2.7(1). On the other hand, we obviously have

(7)

sdepthS(S/Qj) = sdepthS(S/Pj). With these notations, we obtain the follow- ing bounds for sdepthS(I) and sdepthS(S/I).

Corollary 4.6 ([10, Corollary 2.14]).

(1) n− bd1/2c ≥sdepthS(I)≥q1+· · ·+qk−n(k−1).

(2) n−d1 ≥sdepthS(S/I) ≥ min{n−d1, q1−d2, q1+q2−d3 −n, . . . , q1+· · ·+qk−1−dk−n(k−2)}.

Example 4.7 ([10, Example 2.15]). LetI =Q1∩Q2∩Q3 ⊂S :=K[x1, . . . , x7], whereQ1 = (x21, . . . , x25),Q2= (x34, x35, x36) andQ3 = (x36, x6x7, x27). Denote Pj =p

Qj. Note thatq3 = sdepthS(Q3) = sdepthK[x6,x7](Q3∩K[x6, x7]) + 5 = 1 + 5 = 6. Also, since Q1 and Q2 are generated by powers of variables, by [8, Theorem 1.3], q1 = 7− b5/2c = 5 and q2 = 7− b3/2c = 6. According to Corollary 2.14, we have 5 = 7− bd1/2c ≥sdepthS(I)≥q1+q2+q3−14 = 3 and 2 = 7−d1 ≥sdepthS(S/I)≥min{7−d1, q1−d2, q1+q2−d3−7}= min{7−

5,5−3,5 + 6−2−7}= 2. Thus sdepthS(I)∈ {3,4,5}and sdepthS(S/I) = 2.

On the other hand, depthS(S/I)≤min{n−depthS(S/Pj) :j= 1,2,3}= 2.

In particular, we have sdepthS(I)≥depthS(I) and sdepthS(S/I)≥depthS(S/I).

Thus, both I and S/I satisfy the Stanley conjecture. In fact, using CoCoA, we get depthS(S/I) = 2.

In the third section of [10], we give several equivalent form of the Stanley conjecture. We exemplify with the following Proposition.

Proposition 4.8 ([10, Proposition 3.1]). The following assertions are equivalent:

(1)For any integern≥1and any monomial idealI ⊂S=K[x1, . . . , xn], Stanley conjecture holds for I, i.e. sdepthS(I)≥depthS(I).

(2)For any integern≥1and any monomial idealsI, J ⊂S, ifsdepthS(I+ J)≥depthS(I+J), then sdepthS(I)≥depthS(I).

(3)For any integersn, m≥1, any monomial idealI ⊂S=K[x1, . . . , xn], if u1, . . . , um ∈S is a regular sequence on S/I andJ = (u1, . . . , um), then if:

sdepthS(I+J)≥depthS(I+J)⇒sdepthS(I)≥depthS(I).

(4)For any integersn, m≥1, any monomial idealI ⊂S=K[x1, . . . , xn], if u1, . . . , um ∈S is a regular sequence on S/I andJ = (u1, . . . , um), then if:

sdepthS(I+J) = depthS(I+J)⇒sdepthS(I) = depthS(I).

(5) For any integer n≥1, any monomial ideal I ⊂S=K[x1, . . . , xn], if S¯=S[y], then: sdepthS¯(I, y) = depthS(I)⇒sdepthS(I) = depthS(I).

Acknowledgments. The support from grant ID-PCE-2011-1023 of Romanian Min- istry of Education, Research and Innovation is gratefully acknowledged.

(8)

REFERENCES

[1] J. Apel,On a conjecture of R.P. Stanley; Part II – Quotients Modulo Monomial Ideals.

J. Algebraic Combin. 17(2003), 57–74.

[2] C. Biro, D.M. Howard, M.T. Keller, W.T. Trotter and S.J. Young,Interval partitions and Stanley depth. J. Combin. Theory Ser. A117, (2010),4,475–482.

[3] W. Bruns, C. Krattenthaler and J. Uliczka,Stanley decompositions and Hilbert depth in the Koszul complex. J. Commut. Algebra2(2010),3,327–357.

[4] W. Bruns, C. Krattenthaler and J. Uliczka,Hilbert depth of powers of the maximal ideal.

Commutative Algebra and its Connections to Geometry (PASI 2009), Contemporary Mathematics555, Amer. Math. Soc., R.I., 2011, 1–12.

[5] Mircea Cimpoea¸s, Some remarks on the Stanley depth for multigraded modules. Le MathematicheLXIII(2008),II,165–171.

[6] Mircea Cimpoea¸s, Stanley depth of square free Veronese ideals. Preprint (2009).

http://arxiv.org/pdf/0907.1232.

[7] M. Cimpoeas,Stanley depth of monomial ideals with small number of generators. Cent.

Eur. J. Math. 7(2009),4,629–634.

[8] M. Cimpoeas,Stanley depth for monomial complete intersection. Bull. Math. Soc. Sci.

Math. Roumanie (N.S.)51(99)(2008),3,205–211.

[9] M. Cimpoeas,The Stanley conjecture on monomial almost complete intersection ideals.

Bull. Math. Soc. Sci. Math. Roumanie (N.S.)55(103)(2012),1,35–39.

[10] M. Cimpoeas,Several results regarding Stanley depth. Romanian J. Math. Comp. Sci.

2(2012),2,28–40.

[11] J. Herzog, M. Vladoiu and X. Zheng,How to compute the Stanley depth of a monomial ideal. J. Algebra322(2009),6,3151–3169.

[12] Mitchel T. Keller, Yi-Huang Shen, Noah Streib and Stephen J. Young,On the Stanley depth of squarfree Veronese ideals. J. Algebraic Combin. 33(2011),2,313–324.

[13] R. Okazaki,A lower bound of Stanley depth of monomial ideals., J. Commut. Algebra 3(2011),1, 83–88.

[14] A. Popescu,Special Stanley decompositions. Bull. Math. Soc. Sci. Math. Roumanie (N.S.)53(101)(2010),4,363–372.

[15] A. Rauf,Depth and sdepth of multigraded module. Comm. Algebra38(2010),2, 773–

784.

[16] A. Rauf,Stanley Decompositions, Pretty Clean Filtrations and Reductions Modulo Reg- ular Elements. Bull. Math. Soc. Sc. Math. Roumanie (N.S.)50(98)(2007) 347–354.

[17] Y. Shen,Stanley depth of complete intersection monomial ideals and upper-discrete par- titions. J. Algebra321(2009), 1285–1292.

Received 19 March 2013 “Simion Stoilow” Institute of Mathematics, Research unit 5, P.O. Box 1-764,

Bucharest 014700, Romania mircea.cimpoeas@imar.ro

Références

Documents relatifs

The adenovirus carrying the complete cDNA fragment of rat Cx43 inserted in the antisense orientation, fused or not to GFP, strongly (P = 0.0001) inhibited in-vitro maturation of

By treating exhaustion as an economic problem, both on the definition of depletion and on the reasoning method used to evaluate the common solutions advocated,

This approach is based on our preceding works [6, 16] on the Bercovici-Pata bijection (see also [7, 26, 10]). This bijection, that we denote by Λ, is a correspondence between

find an expression profile similar to that observed in M. Still, no straightforward hypothesis can be proposed concerning the absence of DNF1 up-regulation in A. In the case of

Cette dernière espèce a d'ailleurs été trouvée dans l'eau absolument douce ( FRÉMY). Le Polysiphonia opaca, bien que nettement marin, se développe très bien

GANEO AFRICANA (Skrjabin, 1916), (Kaw, 1950) UNE ESPÈCE RARE DE TRÉMATODE PARASITE D’AMPHIBIENS DU TOGO... africana, un Trématode récolté chez un Amphibien (Ranidae

The measured transmission loss peak observed in the partially lined acoustic duct can be explained by the crossing (or avoided crossing) of the lower two Bloch mode attenuations in

Les langues que nous parlons, les villes et les villages que nous habitons, l’écriture, ce que nous savons de la vie et de la course des planètes à travers le ciel,