Loading [MathJax]/extensions/TeX/mathchoice.js
Research article

On Modified Erdős-Ginzburg-Ziv constants of finite abelian groups

  • Received: 01 September 2022 Revised: 22 December 2022 Accepted: 30 December 2022 Published: 09 January 2023
  • MSC : 11B13, 11P70, 05D10

  • Let G be a finite abelian group with exponent exp(G) and S be a sequence with elements of G. We say S is a zero-sum sequence if the sum of the elements in S is the zero element of G. For a positive integer t, let stexp(G)(G) (respectively, stexp(G)(G)) denote the smallest integer such that every sequence (respectively, zero-sum sequence) S over G with |S| contains a zero-sum subsequence of length texp(G). The invariant stexp(G)(G) (respectively, stexp(G)(G)) is called the Generalized Erdős-Ginzburg-Ziv constant (respectively, Modified Erdős-Ginzburg-Ziv constant) of G. In this paper, we discuss the relationship between Generalized Erdős-Ginzburg-Ziv constant and Modified Erdős-Ginzburg-Ziv constant, and determine stexp(G)(G) for some finite abelian groups.

    Citation: Yuting Hu, Jiangtao Peng, Mingrui Wang. On Modified Erdős-Ginzburg-Ziv constants of finite abelian groups[J]. AIMS Mathematics, 2023, 8(3): 6697-6704. doi: 10.3934/math.2023339

    Related Papers:

    [1] Guoqing Wang . A generalization of Kruyswijk-Olson theorem on Davenport constant in commutative semigroups. AIMS Mathematics, 2020, 5(4): 2992-3001. doi: 10.3934/math.2020193
    [2] Guoqing Wang . Lower bound for the Erdős-Burgess constant of finite commutative rings. AIMS Mathematics, 2020, 5(5): 4424-4431. doi: 10.3934/math.2020282
    [3] Xiao-Yu Li, Yu-Lan Wang, Zhi-Yuan Li . Numerical simulation for the fractional-in-space Ginzburg-Landau equation using Fourier spectral method. AIMS Mathematics, 2023, 8(1): 2407-2418. doi: 10.3934/math.2023124
    [4] Narcisse Batangouna . A robust family of exponential attractors for a time semi-discretization of the Ginzburg-Landau equation. AIMS Mathematics, 2022, 7(1): 1399-1415. doi: 10.3934/math.2022082
    [5] Huiqing Wang, Md Nur Alam, Onur Alp İlhan, Gurpreet Singh, Jalil Manafian . New complex wave structures to the complex Ginzburg-Landau model. AIMS Mathematics, 2021, 6(8): 8883-8894. doi: 10.3934/math.2021515
    [6] Guixin Deng, Shuxin Wang . On the Davenport constant of a two-dimensional box $\left[\kern-0.15em\left[ { - 1, 1} \right]\kern-0.15em\right] \times \left[\kern-0.15em\left[ { - m, n} \right]\kern-0.15em\right]$. AIMS Mathematics, 2021, 6(2): 1101-1109. doi: 10.3934/math.2021066
    [7] Mourad S. Semary, M. T. M. Elbarawy, Aisha F. Fareed . Discrete Temimi-Ansari method for solving a class of stochastic nonlinear differential equations. AIMS Mathematics, 2022, 7(4): 5093-5105. doi: 10.3934/math.2022283
    [8] Junling Sun, Xuefeng Han . Existence of Sobolev regular solutions for the incompressible flow of liquid crystals in three dimensions. AIMS Mathematics, 2022, 7(9): 15759-15794. doi: 10.3934/math.2022863
    [9] Rui Wang, Jiangtao Peng . On the inverse problems associated with subsequence sums of zero-sum free sequences over finite abelian groups Ⅱ. AIMS Mathematics, 2021, 6(2): 1706-1714. doi: 10.3934/math.2021101
    [10] Muhammad Imran Asjad, Sheikh Zain Majid, Waqas Ali Faridi, Sayed M. Eldin . Sensitive analysis of soliton solutions of nonlinear Landau-Ginzburg-Higgs equation with generalized projective Riccati method. AIMS Mathematics, 2023, 8(5): 10210-10227. doi: 10.3934/math.2023517
  • Let G be a finite abelian group with exponent exp(G) and S be a sequence with elements of G. We say S is a zero-sum sequence if the sum of the elements in S is the zero element of G. For a positive integer t, let stexp(G)(G) (respectively, stexp(G)(G)) denote the smallest integer such that every sequence (respectively, zero-sum sequence) S over G with |S| contains a zero-sum subsequence of length texp(G). The invariant stexp(G)(G) (respectively, stexp(G)(G)) is called the Generalized Erdős-Ginzburg-Ziv constant (respectively, Modified Erdős-Ginzburg-Ziv constant) of G. In this paper, we discuss the relationship between Generalized Erdős-Ginzburg-Ziv constant and Modified Erdős-Ginzburg-Ziv constant, and determine stexp(G)(G) for some finite abelian groups.



    Let Zn denote the cyclic group of n elements. Let G be a finite abelian group. Up to isomorphism, we can write G as Zn1Znr with 1<n1nr, where nr=exp(G) is the exponent of G. Whenever n1=n2==nr=n, we denote G by Zrn. For convenience, we write gZrn as g=(a1,a2,,ar), where integers ai[0,n1] for every i[1,r].

    Let G be an abelian group and G0G be a subset. We consider sequences over G0 as elements in the free abelian monoid with basis G0. So a sequence S over G0 can be written in the form

    S=g1g=gG0gvg(S),

    where vg(S)N{0} denotes the multiplicity of g in S. We denote |S|=N{0} the length of S, σ(S)=i=1gi=gGvg(S)gG the sum of S. We say the sequence S over G0 is a zero-sum sequence if σ(S)=0G. For an element gG, let g+S denote the sequence (g+g1)(g+g) over g+G0. Let H be a subgroup of G and φ:GG/H be the natural homomorphism. Then φ(S) denotes the sequence φ(g1)φ(g) over G0/H. A sequence T is called a subsequence of S if vg(T)vg(S) for all gG0. Whenever T is a subsequence of S, let ST1 denote the subsequence with T deleted from S. If S1 and S2 are two sequences over G, let S1S2 denote the sequence satisfying that vg(S1S2)=vg(S1)+vg(S2) for all gG0. If S1=S2, we denote S1S2 by S21.

    For a finite abelian group G and a positive integer t, let stexp(G)(G) denote the smallest integer such that every sequence S over G with |S| contains a zero-sum subsequence of length texp(G). For t=1, we abbreviate sexp(G)(G) to s(G), which is called the Erdős-Ginzburg-Ziv constant of G. For t2, the invariant stexp(G)(G) is called the Generalized Erdős-Ginzburg-Ziv constant of G. The classical Erdős-Ginzburg-Ziv Theorem [7], proved in 1961, states that

    s(Zn)=2n1.

    This theorem was viewed as one of the starting points for many problems involving zero-sum sequences over finite abelian groups. These problems occur naturally in various branches of combinatorics and number theory (see [1,3,16] for some classical papers). Since 1960s, the invariant stexp(G)(G) has been studied by many authors (for recent progress see [6,10,12,13,14,15,17,18,19,20]).

    Our main motivation of this paper is the following Modified Erdős-Ginzburg-Ziv constant, which was introduced by Augspurger et al. [2] in 2017.

    Definition 1.1. Let G be an abelian group and G0G be a subset. For a positive integer w, let sw(G0) denote the smallest integer such that every zero-sum sequence S over G0 with |S| contains a zero-sum subsequence of length w. If no such exists, then let sw(G0)=. For w=exp(G), we abbreviate sexp(G)(G0) to s(G0). The invariant sw(G0) is called the Modified Erdős-Ginzburg-Ziv constant of G0.

    There has been lots of results on Generalized Erdős-Ginzburg-Ziv constant, but little is known about Modified Erdős-Ginzburg-Ziv constant.

    For the infinite abelian group Z, let [k,k]={ikik}Z, where k is a positive integer. Augspurger et al. [2] gave some conditions for which sw([k,k]) is finite. They also gave the first results on sw([k,k]) for some positive integers w,k. In 2019, Berger [4] completely determined sw([k,k]) for all positive integers w,k.

    Let G be a finite abelian group and g be an element of G with ord(g)=exp(G). Let w be a positive integer with exp(G)w. Then for every positive integer , both S1=g(g) and S2=g+1((+1)g) are zero-sum sequences over G. But either S1 or S2 contains no zero-sum subsequence of length w. It follows that sw(G) is infinite. So we always consider the case when w=texp(G), where t is a positive integer. In 2009, Berger and Wang [5] determined the invariants stn(Zn) for every positive integer t and s(Z2n).

    By the definition of stexp(G)(G) and stexp(G)(G), we clearly have

    stexp(G)(G)stexp(G)(G) (1.1)

    for every finite abelian group G and every positive integer t.

    Let G be a finite abelian group. In 2017, Zhong (see [2,Lemma 1]) proved that if gcd(s(G)1,exp(G))=1, then s(G)=s(G). In this paper, we can slightly improve this result as

    Theorem 1.2. Let G be a finite abelian group. If gcd(stexp(G)(G)i,exp(G))=1 for some positive integer i, then stexp(G)(G)>stexp(G)(G)i. Moreover, if gcd(stexp(G)(G)1,exp(G))=1, then stexp(G)(G)=stexp(G)(G).

    By Theorem 1.2, we are able to determine stexp(G)(G) for the following cases.

    Theorem 1.3. (1) Let G=Z2n. If t2 and gcd(n,3)=1, then stn(G)=stn(G)=tn+2n2.

    (2) Let G=ZmZn with 1mn. If gcd(2mi,n)=1 for some integer i4, then s(G)2m+2n(i1). Moveover, if i=4, then s(G)=s(G)=2m+2n3.

    (3) Let G=Zd2, where d is even. If 2t>d, then s2t(G)=s2t(G)=2t+d.

    We also find some cases such that stexp(G)(G)stexp(G)(G). We need the following definition associated with s(G).

    Definition 1.4. [8,Definition 1.1] A pair (n,d) of positive integers is said to have Property D if (n1)(s(Zdn)1) and every sequence S over Zdn of length s(Zdn)1 having no zero-sum subsequence of length n is of the form Tn1, where T is a sequence over Zdn with |T|=(s(Zdn)1)/(n1).

    Let G=Zrn with n3 and r2. Suppose s(G)=c(n1)+1 for some positive integer c and (n,r) has Property D. Zhong (see [2,Lemma 1]) proved that if gcd(s(G)1,n)=c, then s(G)<s(G). In this paper, we obtain that

    Theorem 1.5. (1) Let n,d,k1 be positive integers and n>1. Suppose s(Zdnk)=c(nk1)+1 for every positive integer k, where c is a constant depending only on n and d. If s(Zdnk1)s(Zdnk1)i for some positive integer i, then s(Zdnk)s(Zdnk)i for every integer kk1.

    (2) Let G=ZmZn with 1mn and H=ZdZd with dm. If s(H)s(H)i for some positive integer i, then s(G)s(G)i.

    By using Theorem 1.2 and Theorem 1.5, we obtain the following results.

    Theorem 1.6. Let k,d be two positive integers.

    (1) If k<d, then s(Zd2k)=s(Zd2k)=2d(2k1)+1.

    (2) If kd, then s(Zd2k)=s(Zd2k)1=2d(2k1).

    (3) If k2, then s(Z33k)=s(Z33k)1=9(3k1).

    We remark that the first two results in Theorem 1.6 confirm a conjecture of Berger and Wang (see [5,Conjecture 4.2]).

    This paper is organized as follows. In Section 2, we deal with Theorems 1.2 and 1.3. In Section 3, we prove Theorems 1.5 and 1.6. In the final section, some additional results are given.

    In this section, we deal with some cases when stexp(G)(G)=stexp(G)(G).

    Proof of Theorem 1.2. Suppose gcd(stexp(G)(G)i,exp(G))=1 for some positive integer i. We need to find a zero-sum sequence of length stexp(G)(G)i, which contains no zero-sum subsequence of length texp(G). By the definition of stexp(G)(G), there exists a sequence S with |S|=stexp(G)(G)i and S has no zero-sum subsequence of length texp(G). Assume that σ(S)=hG and let xN be an integer such that

    (\mathtt{s}_{t\exp(G)}(G)-i)x\equiv 1 \pmod {\exp(G)}.

    Such integer x exists because \gcd(\mathtt{s}_{t\exp(G)}(G)-i, \exp(G)) = 1 . Let

    S' = -(xh)+S.

    Since \sigma(S') = \sigma(S)-(\mathtt{s}_{t\exp(G)}(G)-i)xh = 0, we obtain that S' is a zero-sum sequence of length \mathtt{s}_{t\exp(G)}(G)-i . Next we prove that S' has no zero-sum subsequence of length t\exp(G) . Assume to the contrary that T' is a zero-sum subsequence of S' with |T'| = t\exp(G). Then \sigma(T')+t\exp(G)(xh) = 0 . It follows that T = (xh)+T' is a zero-sum subsequence of S with |T| = t\exp(G), yielding a contradiction to the definition of S . So S' has no zero-sum subsequence of length t\exp(G) . It follows that \mathtt{s}'_{t\exp(G)}(G) > \mathtt{s}_{t\exp(G)}(G)-i , and we are done.

    Moreover, if \gcd(\mathtt{s}_{t\exp(G)}(G)-1, \exp(G)) = 1 , then \mathtt{s}'_{t\exp(G)}(G) > \mathtt{s}_{t\exp(G)}(G)-1 . By (1.1), we obtain that \mathtt{s}'_{t\exp(G)}(G)\leq \mathtt{s}_{t\exp(G)}(G) . Therefore, \mathtt{s}'_{t\exp(G)}(G) = \mathtt{s}_{t\exp(G)}(G) .

    We need the following lemmas to prove Theorem 1.3.

    Lemma 2.1. [9,Proposition 4.1] Suppose t\geq2 , then \mathtt{s}_{tn}(\mathbb{Z}_{n}^2) = tn+2n-2 .

    Lemma 2.2. [11,Theorem 5.8.3] Let G = \mathbb{Z}_m\oplus \mathbb{Z}_n with 1\leq m \mid n . Then \mathtt{s}(G) = 2m+2n-3 .

    Lemma 2.3. [20,Theorem 5.9] \mathtt{s}_{2t}(\mathbb{Z}_2^{d}) = 2t+d for d < 2t .

    Proof of Theorem 1.3. (1). Let G = \mathbb{Z}_{n}^2 . Suppose t \geq 2 . By Lemma 2.1, we have that \mathtt{s}_{tn}(G) = tn+2n-2 . Since \gcd(n, 3) = 1 , we infer that \gcd(\mathtt{s}_{tn}(G)-1, n) = 1 . It follows from Theorem 1.2 that \mathtt{s}'_{tn}(G) = \mathtt{s}_{tn}(G) = tn+2n-2 .

    (2). Let G = \mathbb{Z}_m\oplus \mathbb{Z}_n . By Lemma 2.2, we have that \mathtt{s}(G) = 2m+2n-3 . Since \gcd(2m-i, n) = 1 for some integer i\geq 4 , we infer that \gcd(\mathtt{s}(G)-(i-3), \exp(G)) = 1 . By Theorem 1.2, we obtain that \mathtt{s}'(G)\geq\mathtt{s}(G)-(i-3)+1 = 2m+2n-(i-1) .

    If i = 4 , we infer that \gcd(\mathtt{s}(G)-1, \exp(G)) = 1 . By Theorem 1.2, we obtain that \mathtt{s}'(G) = \mathtt{s}(G) = 2m+2n-3 .

    (3). Let G = \mathbb{Z}_2^d . Since d < 2t , by Lemma 2.3, we have that \mathtt{s}_{2t}(G) = 2t+d . Since d is even, we obtain that \gcd(2t+d-1, 2) = 1 , i.e. \gcd(\mathtt{s}_{2t}(G)-1, \exp(G)) = 1 . By Theorem 1.2, we obtain that \mathtt{s}_{2t}'(G) = \mathtt{s}_{2t}(G) = 2t+d .

    In this section, we study some cases when \mathtt{s}'_{t\exp(G)}(G)\neq\mathtt{s}_{t\exp(G)}(G) .

    Proof of Theorem 1.5. (1). We proceed by induction on k . If k = k_1 , then by the hypothesis of the theorem we have \mathtt{s}'(\mathbb{Z}_{n^{k_1}}^d)\leq \mathtt{s}(\mathbb{Z}_{n^{k_1}}^d)-i .

    Suppose that the theorem is true for k-1 \geq k_1 . We will show that \mathtt{s}'(\mathbb{Z}_{n^k}^d)\leq \mathtt{s}(\mathbb{Z}_{n^k}^d)-i . Let S be a zero-sum sequence over \mathbb{Z}_{n^k}^d with |S| = \mathtt{s}(\mathbb{Z}_{n^k}^d)-i = c(n^k-1)+1-i . It suffices to show that S contains a zero-sum subsequence of length n^k . Let

    \phi:\ \mathbb{Z}_{n^k}^d\ \rightarrow \ \mathbb{Z}_{n^{k-1}}^d

    be the natural homomorphism with \ker \phi \cong \mathbb{Z}_{n}^d . Then \phi(S) is a zero-sum sequence over \mathbb{Z}_{n^{k-1}}^d with |\phi(S)| = c(n^k-1)+1-i = c(n-1)\times n^{k-1}+c(n^{k-1}-1)+1-i > c(n^{k-1}-1)+1-i . By the assumption, we have \mathtt{s}'(\mathbb{Z}_{n^{k-1}}^d)\leq c(n^{k-1}-1)+1-i . So we can find a subsequence S_1 of S such that \phi(S_1) is a zero-sum sequence over \mathbb{Z}_{n^{k-1}}^d of length n^{k-1} . It follows that \phi(SS_1^{-1}) is a zero-sum sequence over \mathbb{Z}_{n^{k-1}}^d with |\phi(SS_1^{-1})| = c(n^k-1)+1-i-n^{k-1} = (c(n-1)-1)\times n^{k-1}+c(n^{k-1}-1)+1-i > c(n^{k-1}-1)+1-i (n > 1) . Then we can find a subsequence S_2 of SS_1^{-1} such that \phi(S_2) is a zero-sum sequence over \mathbb{Z}_{n^{k-1}}^d of length n^{k-1} . Repeating the above procedure, one can find c(n-1)+1 disjoint zero-sum subsequences \phi(S_1), \ldots, \phi(S_{c(n-1)+1}) of \phi(S) such that |\phi(S_1)| = \cdots = |\phi(S_{c(n-1)+1})| = n^{k-1} . Since \phi(\sigma(S_i)) = \sigma(\phi(S_i)) = \overline{0}\in\mathbb{Z}_{n^{k-1}}^d , we infer that \sigma (S_i)\in \ker\phi \cong \mathbb{Z}_n^d for i = 1, \ldots, c(n-1)+1 . Let

    T = \sigma (S_1)\cdot\ldots\cdot\sigma (S_{c(n-1)+1}).

    Then T is a sequence over \ker\phi \cong \mathbb{Z}_n^d . By the hypothesis of the theorem, we have \mathtt{s}(\mathbb{Z}_n^d) = c(n-1)+1 . So T has a zero-sum subsequence of length n . Without loss of generality we may assume that \sigma (S_1)\sigma (S_2)\cdot\ldots\cdot\sigma (S_n) is a zero-sum subsequence of T over \ker\phi \cong \mathbb{Z}_n^d . Since \sigma (S_1)+\sigma (S_2)+\cdots+\sigma (S_n) = 0 \in \ker\phi \subseteq \mathbb{Z}_{n^k}^d and |S_1S_2\cdot\ldots\cdot S_n| = |\phi(S_1)\phi(S_2)\cdot\ldots\cdot\phi (S_n)| = n^k , we obtain that S_1S_2\cdot\ldots\cdot S_n is a zero-sum subsequence of S of length n^k . It follows that \mathtt{s}'(\mathbb{Z}_{n^k}^d)\leq \mathtt{s}(\mathbb{Z}_{n^k}^d)-i , and we are done.

    (2). Let G = \mathbb{Z}_m\oplus \mathbb{Z}_n with 1\leq m \mid n and H = \mathbb{Z}_d\oplus \mathbb{Z}_d with d\mid m . By Lemma 2.2, we have \mathtt{s}(G) = 2m+2n-3 . Let S be a zero-sum sequence over G with |S| = \mathtt{s}(G)-i = 2m+2n-3-i . It suffices to show that S contains a zero-sum subsequence of length n . Let

    \phi:\ G\ \rightarrow \ H

    be the natural homomorphism with \ker \phi \cong G/H = \mathbb{Z}_{\frac{m}{d}}\oplus \mathbb{Z}_{\frac{n}{d}} . By the assumption, we have \mathtt{s}'(H)\leq 4d-3-i . Since \phi(S) is a zero-sum sequence over H with |\phi(S)| = 2m+2n-3-i = (2\frac{m}{d}+2\frac{n}{d}-4)d+4d-3-i , similar with (1), we can find 2\frac{m}{d}+2\frac{n}{d}-3 disjoint zero-sum subsequences \phi(S_1), \ldots, \phi(S_{2\frac{m}{d}+2\frac{n}{d}-3}) such that |\phi(S_1)| = \cdots = |\phi(S_{2\frac{m}{d}+2\frac{n}{d}-3})| = d . By Lemma 2.2, we have \mathtt{s}(G/H) = 2\frac{m}{d}+2\frac{n}{d}-3 . So we can prove that S has a zero-sum subsequence of length n . It follows that \mathtt{s}'(G)\leq \mathtt{s}(G)-i , and we are done.

    In 2003 , Gao and Thangadurai proved the following result.

    Lemma 3.1. [8,Corollary 1.1] The pairs (2^k, d) , (3^k, 3) have Property D for any positive integers k and d .

    We need the following results.

    Lemma 3.2. (1) \mathtt{s}(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1)+1 for every k\geq 1 and d\geq 1 . (Harborth [16])

    (2) \mathtt{s}(\mathbb{Z}_{3^k}^3) = 9(3^k-1)+1 for every k\geq 1 . (Gao and Thangadurai [8])

    Now we are in a position to prove Theorem 1.6.

    Proof of Theorem 1.6. (1). We will prove \mathtt{s}'(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1) +1 for k \leq d-1 . By (1.1) and Lemma 3.2.(1), we have \mathtt{s}'(\mathbb{Z}_{2^k}^d) \leq \mathtt{s}(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1)+1. Let

    W = \prod\limits_{a_1, a_2, \ldots, a_d\in \{0, 1\}}(a_1, a_2, \ldots, a_d)^{2^k-1}.

    Since k \leq d-1 , it is easy to check that W is a zero-sum sequence of length 2^d(2^k-1) , and W contains no zero-sum subsequence of length 2^k . It follows that \mathtt{s}'(\mathbb{Z}_{2^k}^d) \geq 2^d(2^k-1)+1 . Therefore, \mathtt{s}'(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1)+1 for k \leq d-1 .

    (2). Assume that k\geq d . By (1.1) and Lemma 3.2.(1), we have \mathtt{s}'(\mathbb{Z}_{2^k}^d) \leq \mathtt{s}(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1)+1. We first prove that \mathtt{s}'(\mathbb{Z}_{2^k}^k)\leq 2^k(2^k-1) . Let S be a zero-sum sequence over \mathbb{Z}_{2^k}^k with |S| = 2^k(2^k-1) . We need to prove that S contains a zero-sum subsequence of length 2^k . Assume to the contrary that S contains no zero-sum subsequence of length 2^k . Note that |S| = \mathtt{s}(\mathbb{Z}_{2^k}^k)-1 . By Lemma 3.1, we obtain that the pair (2^k, k) has Property D. Therefore, S is of form T^{2^k-1} for some sequence T with |T| = \frac{s(\mathbb{Z}_{2^k}^k)-1}{2^k-1} = 2^k . Since 0 = \sigma(S) = (2^k-1) \sigma(T) and \gcd(2^k-1, 2^k) = 1 , we infer that \sigma(T) = 0 . Therefore, T is a zero-sum subsequence of S with |T| = 2^k , yielding a contradiction. So S contains a zero-sum subsequence of length 2^k , and this proves that

    \mathtt{s}'(\mathbb{Z}_{2^k}^k)\leq 2^k(2^k-1) = \mathtt{s}(\mathbb{Z}_{2^k}^k)-1.

    By Theorem 1.5, we have that \mathtt{s}'(\mathbb{Z}_{2^k}^d)\leq \mathtt{s}(\mathbb{Z}_{2^k}^d)-1 = 2^d(2^k-1) for every k\geq d . Since \gcd(2^d(2^k-1)-1, 2^k) = \gcd(\mathtt{s}(\mathbb{Z}_{2^k}^d)-2, \exp(\mathbb{Z}_{2^k}^d)) = 1 , by Theorem 1.2, we obtain that

    \mathtt{s}'(\mathbb{Z}_{2^k}^d) > \mathtt{s}(\mathbb{Z}_{2^k}^d)-2 = 2^d(2^k-1)-1.

    Therefore, \mathtt{s}'(\mathbb{Z}_{2^k}^d) = 2^d(2^k-1) for every k \geq d .

    (3). By Lemma 3.2.(2), we have \mathtt{s}(\mathbb{Z}_{3^k}^3) = 9(3^k-1)+1 . Since (3^k, 3) has Property D, by using a similar argument as in (2), we can prove that \mathtt{s}'(\mathbb{Z}_{3^k}^3) = 9(3^k-1) for every integer k\geq 2 .

    In this section, we give some additional results. By (1.1), we have \mathtt{s}'_{t\exp(G)}(G)\leq \mathtt{s}_{t\exp(G)}(G) for every finite abelian group G and every positive integer t . By Theorem 1.6, the equality \mathtt{s}'_{t\exp(G)}(G) = \mathtt{s}_{t\exp(G)}(G) does not always hold. A natural problem asks that

    Problem 4.1. When do we have \mathtt{s}'_{t\exp(G)}(G) = \mathtt{s}_{t\exp(G)}(G) ?

    By Theorem 1.2, if \gcd(\mathtt{s}_{t\exp(G)}(G)-1, \exp(G)) = 1 , then \mathtt{s}'_{t\exp(G)}(G) = \mathtt{s}_{t\exp(G)}(G) . However, \gcd(\mathtt{s}_{t\exp(G)}(G)-1, \exp(G)) = 1 is not necessary for \mathtt{s}'_{t\exp(G)}(G) = \mathtt{s}_{t\exp(G)}(G) . In this paper, we are able to prove the following result.

    Theorem 4.2. Let G = \mathbb{Z}_2^{2t} . Then \mathtt{s}'_{2t}(G) = \mathtt{s}_{2t}(G) = 4t+1 .

    In order to prove Theorem 4.2, we need the following result due to Sidorenko [20].

    Lemma 4.3. [20,Theorem 5.10] \mathtt{s}_{2t}(\mathbb{Z}_2^{2t}) = 4t+1 .

    Proof of Theorem 4.2. By (1.1) and Lemma 4.3, we have

    \mathtt{s}'_{2t}(\mathbb{Z}_2^{2t}) \leq \mathtt{s}_{2t}(\mathbb{Z}_2^{2t}) = 4t+1.

    Next we need to prove \mathtt{s}'_{2t}(\mathbb{Z}_2^{2t})\geq 4t+1 . We denote {\bf e}_j = (a_1, a_2, \ldots, a_{2t}) for every j = 1, 2, \ldots, 2t , where a_j = 1 and a_i = 0 for every i\neq j . Let

    T = {\bf 0}^{2t-1}{\bf e}_1\cdot{\bf e}_2\cdot\ldots \cdot {\bf e}_{2t}\cdot({\bf e}_1+{\bf e}_2+\cdots+{\bf e}_{2t}).

    Then T is a zero-sum sequence with |T| = 4t and T contains no zero-sum subsequence of length 2t . Hence, \mathtt{s}'_{2t}(\mathbb{Z}_2^{2t})\geq 4t+1 . Therefore, \mathtt{s}'_{2t}(\mathbb{Z}_2^{2t}) = 4t+1 .

    Among other results, we can also determine some cases when \mathtt{s}'_{t\exp(G)}(G) < \mathtt{s}_{t\exp(G)}(G) holds.

    Theorem 4.4. (1) If G = \mathbb{Z}_{n} with \gcd(n, 2)\neq 1 , then \mathtt{s}'_{tn}(G) < \mathtt{s}_{tn}(G) for every integer t\geq 1 .

    (2) If G = \mathbb{Z}_{n}^2 with 4\mid n , then \mathtt{s}'(G) < \mathtt{s}(G) .

    (3) If G = \mathbb{Z}_m\oplus \mathbb{Z}_n with 4\mid m \mid n , then \mathtt{s}'(G) < \mathtt{s}(G) .

    In order to prove Theorem 4.4, we need the following results.

    Lemma 4.5. [5,Theorem 1.3] \mathtt{s}'_{tn}(\mathbb{Z}_{n}) = (t+1)n-\ell+1 for every integer t \geq 1 , where \ell is the smallest integer such that \ell \nmid n .

    Lemma 4.6. [5,Theorem 1.4] \mathtt{s}'(\mathbb{Z}_{n}^2) = 4n-\ell +1 , where \ell is the smallest integer such that \ell \geq 4 and \ell \nmid n .

    The following result is a consequence of the Erdős-Ginzburg-Ziv Theorem.

    Lemma 4.7. Let t, n be two positive integers. Then \mathtt{s}_{tn}(\mathbb{Z}_{n}) = (t+1)n-1 .

    Proof of Theorem 4.4. (1). By Lemma 4.5, we infer that \mathtt{s}'_{tn}(\mathbb{Z}_{n}) = (t+1)n-\ell+1 for every integer t \geq 1 , where \ell is the smallest integer such that \ell \nmid n . Since \gcd(n, 2)\neq 1 , we infer that \ell \geq 3 . It follows that \mathtt{s}'_{tn}(\mathbb{Z}_{n})\leq (t+1)n-2 . By Lemma 4.7, we have that \mathtt{s}'_{tn}(\mathbb{Z}_{n}) < \mathtt{s}_{tn}(\mathbb{Z}_{n}) .

    (2). By Lemma 4.6, we infer that \mathtt{s}'(\mathbb{Z}_{n}^2) = 4n-\ell +1 , where \ell is the smallest integer such that \ell \geq 4 and \ell \nmid n . Since 4 \mid n , we infer that \ell \geq 5 . It follows that \mathtt{s}'(\mathbb{Z}_{n}^2) \leq 4n-4 . By Lemma 2.2, we get that \mathtt{s}'(\mathbb{Z}_{n}^2) < \mathtt{s}(\mathbb{Z}_{n}^2). (3). Let H = \mathbb{Z}_{4}^2 . By (2), we obtain that \mathtt{s}'(H)\leq \mathtt{s}(H)-1 . By Theorem 1.5, we infer that \mathtt{s}'(\mathbb{Z}_m\oplus \mathbb{Z}_n) < \mathtt{s}(\mathbb{Z}_m\oplus \mathbb{Z}_n) .

    We would like to thank the referees for their very carefully reading and very useful suggestions. This research was supported in part by the Fundamental Research Funds for the Central Universities (No. 3122019152).

    All authors declare no conflicts of interest in this paper.



    [1] N. Alon, S. Friedland, G. Kalai, Regular subgraphs of almost regular graphs, J. Combin. Theory Ser. B, 37 (1984), 79–91. https://doi.org/10.1016/0095-8956(84)90047-9 doi: 10.1016/0095-8956(84)90047-9
    [2] C. Augspurger, M. Minter, K. Shoukry, P. Sissokho, K. Voss, Avoiding zero-sum sequences of prescribed length over the integers, Integers, 17 (2017), Paper No. A18, 13pp.
    [3] A. Bialostocki, P. Dierker, On the Erdős-Ginzburg-Ziv theorem and the Ramsey numbers for stars and matchings, Discrete Math., 110 (1992), 1–8. https://doi.org/10.1016/0012-365X(92)90695-C doi: 10.1016/0012-365X(92)90695-C
    [4] A. Berger, An analogue of the Erdős-Ginzburg-Ziv Theorem over \mathbb{Z}, Discrete Math., 342 (2019), 815–820. https://doi.org/10.1016/j.disc.2018.11.018 doi: 10.1016/j.disc.2018.11.018
    [5] A. Berger, D. Wang, Modified Erdős-Ginzburg-Ziv constants for \mathbb{Z}/n\mathbb{Z} and (\mathbb{Z}/n\mathbb{Z})^2, Discrete Math., 342 (2019), 1113–1116. https://doi.org/10.1016/j.disc.2018.12.024 doi: 10.1016/j.disc.2018.12.024
    [6] J. Bitz, S Griffith, X. He, Exponential lower bounds on the generalized Erdős-Ginzburg-Ziv constant, Discrete Math., 343 (2020), 112083, 4pp. https://doi.org/10.1016/j.disc.2020.112083 doi: 10.1016/j.disc.2020.112083
    [7] P. Erdős, A. Ginzburg, A. Ziv, Theorem in the additive number theory, Bull. Res. Council Israel, 10 (1961), 41–43.
    [8] W. Gao, R. Thangadurai, On the structure of sequences with forbidden zero-sum subsequences, Colloq. Math., 98 (2003), 213–222. https://doi.org/10.4064/cm98-2-7 doi: 10.4064/cm98-2-7
    [9] W. Gao, D. Han, J. Peng, F. Sun, On zero-sum subsequences of length k\exp(G), J. Combin. Theory Ser. A, 125 (2014), 240–253. https://doi.org/10.1016/j.jcta.2014.03.006 doi: 10.1016/j.jcta.2014.03.006
    [10] W. Gao, S. Hong, J. Peng, On zero-sum subsequences of length k\exp(G) II, J. Combin. Theory Ser. A, 187 (2022), 105563. https://doi.org/10.1016/j.jcta.2021.105563 doi: 10.1016/j.jcta.2021.105563
    [11] A. Geroldinger, F. Halter-Koch, Non-Unique Factorizations. Algebraic, Combinatorial and Analytic Theory, Pure and Applied Mathematics, vol. 278. Chapman & Hall/CRC, Boca Raton, FL, 2006. https://doi.org/10.1201/9781420003208
    [12] B. Girard, W.A. Schmid, Direct zero-sum problems for certain groups of rank three, J. Number Theory, 197 (2019), 297–316. https://doi.org/10.1016/j.jnt.2018.08.016 doi: 10.1016/j.jnt.2018.08.016
    [13] B. Girard, W.A. Schmid, Inverse zero-sum problems for certain groups of rank three, Acta Math. Hungar., 160 (2020), 229–247. https://doi.org/10.1007/s10474-019-00983-w doi: 10.1007/s10474-019-00983-w
    [14] D. Han, H. Zhang, On generalized Erdős-Ginzburg-Ziv constants of C_n^{r}, Discrete Math., 342 (2019), 1117–1127. https://doi.org/10.1016/j.disc.2018.12.018 doi: 10.1016/j.disc.2018.12.018
    [15] D. Han, H. Zhang, Zero-sum invariants on finite abelian groups with large exponent, Discrete Math., 342 (2019), 111617, 7pp. https://doi.org/10.1016/j.disc.2019.111617 doi: 10.1016/j.disc.2019.111617
    [16] H. Harborth, Ein Extremalproblem für Gitterpunkte, J. Reine Angew. Math., 262 (1973), 356–360. https://doi.org/10.1515/crll.1973.262-263.356 doi: 10.1515/crll.1973.262-263.356
    [17] G. Hegedűs, An improved exponential upper bound for the Erdős-Ginzburg-Ziv constant, Integers, 20 (2020), Paper No. A22, 5pp.
    [18] E. Naslund, Exponential bounds for the Erdős-Ginzburg-Ziv constant, J. Combin. Theory Ser. A, 174 (2020), 105185, 19pp. https://doi.org/10.1016/j.jcta.2019.105185 doi: 10.1016/j.jcta.2019.105185
    [19] J. Oh, Q. Zhong, On Erdős-Ginzburg-Ziv inverse theorems for dihedral and dicyclic groups, Israel J. Math., 238 (2020), 715–743. https://doi.org/10.1007/s11856-020-2036-6 doi: 10.1007/s11856-020-2036-6
    [20] A. Sidorenko, On generalized Erdős-Ginzburg-Ziv constants for Z_2^d, J. Combin. Theory Ser. A, 174 (2020), 150254, 20pp. https://doi.org/10.1016/j.jcta.2020.105254 doi: 10.1016/j.jcta.2020.105254
  • This article has been cited by:

    1. Alexander Sidorenko, Modified Erdős–Ginzburg–Ziv Constants for \mathbb {Z}_2^{d}, 2023, 39, 0911-0119, 10.1007/s00373-023-02709-w
  • Reader Comments
  • © 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1580) PDF downloads(58) Cited by(1)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog