Loading [MathJax]/jax/element/mml/optable/MathOperators.js
Research article

Notes on Hong's conjecture on nonsingularity of power LCM matrices

  • Received: 17 January 2022 Revised: 08 March 2022 Accepted: 10 March 2022 Published: 23 March 2022
  • MSC : Primary 11C20; Secondary 11A05, 15B36

  • Let a,n be positive integers and S={x1,...,xn} be a set of n distinct positive integers. The set S is said to be gcd (resp. lcm) closed if gcd(xi,xj)S (resp. [xi,xj]S) for all integers i,j with 1i,jn. We denote by (Sa) (resp. [Sa]) the n×n matrix having the ath power of the greatest common divisor (resp. the least common multiple) of xi and xj as its (i,j)-entry. In this paper, we mainly show that for any positive integer a with a2, the power LCM matrix [Sa] defined on a certain class of gcd-closed (resp. lcm-closed) sets S is nonsingular. This provides evidences to a conjecture raised by Shaofang Hong in 2002.

    Citation: Guangyan Zhu, Kaimin Cheng, Wei Zhao. Notes on Hong's conjecture on nonsingularity of power LCM matrices[J]. AIMS Mathematics, 2022, 7(6): 10276-10285. doi: 10.3934/math.2022572

    Related Papers:

    [1] Guangyan Zhu, Mao Li, Xiaofan Xu . New results on the divisibility of power GCD and power LCM matrices. AIMS Mathematics, 2022, 7(10): 18239-18252. doi: 10.3934/math.20221003
    [2] Zhengjin Bu, Zhefeng Xu . Asymptotic formulas for generalized gcd-sum and lcm-sum functions over r-regular integers (mod nr). AIMS Mathematics, 2021, 6(12): 13157-13169. doi: 10.3934/math.2021760
    [3] Panpan Liu, Haifeng Sang, Min Li, Guorui Huang, He Niu . New criteria for nonsingular H-matrices. AIMS Mathematics, 2023, 8(8): 17484-17502. doi: 10.3934/math.2023893
    [4] Lanlan Liu, Yuxue Zhu, Feng Wang, Yuanjie Geng . Infinity norm bounds for the inverse of SDD+1 matrices with applications. AIMS Mathematics, 2024, 9(8): 21294-21320. doi: 10.3934/math.20241034
    [5] Long Chen, Shaofang Hong . Divisibility among determinants of power matrices associated with integer-valued arithmetic functions. AIMS Mathematics, 2020, 5(3): 1946-1959. doi: 10.3934/math.2020130
    [6] Zongbing Lin, Shaofang Hong . The least common multiple of consecutive terms in a cubic progression. AIMS Mathematics, 2020, 5(3): 1757-1778. doi: 10.3934/math.2020119
    [7] Yang Chen, Kezheng Zuo, Zhimei Fu . New characterizations of the generalized Moore-Penrose inverse of matrices. AIMS Mathematics, 2022, 7(3): 4359-4375. doi: 10.3934/math.2022242
    [8] Hongjian Li, Pingzhi Yuan . Fermat's and Catalan's equations over M2(Z). AIMS Mathematics, 2023, 8(1): 977-996. doi: 10.3934/math.2023047
    [9] Wanlin Jiang, Kezheng Zuo . Further characterizations of the m-weak group inverse of a complex matrix. AIMS Mathematics, 2022, 7(9): 17369-17392. doi: 10.3934/math.2022957
    [10] Fengxia Jin, Feng Wang, Kun Zhao, Huatao Chen, Juan L.G. Guirao . The method of judging satisfactory consistency of linguistic judgment matrix based on adjacency matrix and 3-loop matrix. AIMS Mathematics, 2024, 9(7): 18944-18967. doi: 10.3934/math.2024922
  • Let a,n be positive integers and S={x1,...,xn} be a set of n distinct positive integers. The set S is said to be gcd (resp. lcm) closed if gcd(xi,xj)S (resp. [xi,xj]S) for all integers i,j with 1i,jn. We denote by (Sa) (resp. [Sa]) the n×n matrix having the ath power of the greatest common divisor (resp. the least common multiple) of xi and xj as its (i,j)-entry. In this paper, we mainly show that for any positive integer a with a2, the power LCM matrix [Sa] defined on a certain class of gcd-closed (resp. lcm-closed) sets S is nonsingular. This provides evidences to a conjecture raised by Shaofang Hong in 2002.



    Throughout, a,n stand for positive integers. Let S={x1,...,xn} be a set of n distinct positive integers, where x1<...<xn. We denote by (x,y) (resp. [x,y]) the greatest common divisor (resp. least common multiple) of integers x and y. Let (f((xi,xj))) (abbreviated by (f(S))) be the n×n matrix whose (i,j)-entry is that the arithmetic function f evaluates at (xi,xj). Let (f([xi,xj])) (abbreviated by (f[S])) be the n×n matrix whose (i,j)-entry is that f evaluates at [xi,xj]. Let ξa be the arithmetic function defined by ξa(x)=xa for any positive integer x. The n×n matrix (ξa((xi,xj))) (abbreviated by (Sa)) and (ξa([xi,xj])) (abbreviated by [Sa]) are called power GCD matrix on S and power LCM matrix on S, respectively. When a=1, they are simply called GCD matrix and LCM matrix. In 1875, Smith [15] showed that

    det(f((i,j)))=nk=1(fμ)(k), (1.1)

    where μ is the Möbius function and fμ is the Dirichlet convolution of the arithmetic function f and μ. Since then, lots of generalizations of Smith's determinant (1.1) and related results have been published (see, for example, [1,2,3,4,5,6,7,8,9,10,11,12,13,14,16,17,18]).

    The study of power GCD matrices was initiated by Bourque and Ligh [2]. They showed that every power GCD matrix is positive definite. A set S is called factor closed (FC) if the conditions xS and d|x imply that dS. We say that the set S is gcd (resp. lcm) closed if (xi,xj)S (resp. [xi,xj]S) for all integers i,j with 1i,jn. Evidently, any FC set is gcd closed but not conversely. We know [15] that (S) and [S] are nonsingular when S is FC. In [1], Bourque and Ligh obtained a formula for the inverse of these matrices on FC sets. Furthermore, they conjectured that [S] is nonsingular if S is gcd closed. In [12], Haukkanen, Wang and Sillanp¨a¨a presented a counterexample to disprove the Bourque-Ligh conjecture [1].

    Definition 1.1. [4,10] Let r be an integer with 1rn1. The set S is 0-fold gcd closed if S is gcd closed. The set S is r-fold gcd closed if there is a divisor chain RS with |R|=r such that max(R)|min(SR) and the set SR is gcd closed.

    Hong [4] proved the Bourque-Ligh conjecture if n5, and also if n6 and S is (n5)-fold. He [6] proved this conjecture for n7 and disproved for n8. Further, he proved it under certain assumptions [5], and proposed [7] the following conjecture.

    Conjecture 1.1. [7] There is a positiveinteger k(a) depending only on a, such that [Sa] is nonsingularif nk(a) and S is gcd closed. But for anyinteger nk(a)+1, there is a gcd-closed set S such that [Sa] is singular.

    Hong [7] noted that k(a)7 for any a2. He also [8] showed that [Sa] is nonsingular if S is gcd closed and all its elements have at most two distinct prime factors. We denote by lcm(A) (resp. gcd(A)) the least common multiple (resp. greatest common divisor) of all elements of A, where A is a set of finite distinct positive integers.

    Definition 1.2. [10] Let m=lcm(S). The reciprocal set of S is

    mS1={mx1,...,mxn}.

    Definition 1.3. [14] Let r be an integer with 1rn1. The set S is r-fold lcm closed if mS1 is r-fold gcd closed.

    Definition 1.4. [10,14] The n×n matrix (1Sa) having the reciprocal ath power of (xi,xj) as its (i,j)-entry is called reciprocal power GCD matrix defined on S. When a=1, it is simply called reciprocal GCD matrix on S, and denoted by (1S).

    Hong [10] proved that if n7 and S is gcd (resp. lcm) closed, then [S] and (1S) are nonsingular, which also holds if n8 and S is (n7)-fold gcd (resp. lcm) closed. Hong, Shum and Sun [11] proved that [Sa] is nonsingular if every element of a gcd-closed set S is of the form pqr, or p2qr, or p3qr except for the case a=1 and 270,520S, where p,q,r are distinct primes. They also showed that [Sa] is nonsingular if S is a gcd-closed set satisfying xi<180 for all integers i with 1in. Cao [3] developed Hong's method and proved that k(a)8 for all a2. Li [14] showed that if n7 and S is gcd (resp. lcm) closed, then ([xi,xj]e) and (1Se) are nonsingular for any real number e with e1, and also if n8 and S is (n7)-fold gcd (resp. lcm) closed. Wan, Hu and Tan [16] extended Hong's results [8] and [11] by showing that [Sa] is nonsingular when S is gcd closed such that every element of S contains at most two distinct prime factors or is of the form plqr with p,q,r being distinct primes and the positive integer l satisfying 1l4 except for the case a=1 and 270,520,810,1040S. Let x,yS with x<y. If x|y and the conditions x|d|y and dS imply that d{x,y}, then we say that x is a greatest-type divisor of y in S. For xS, GS(x) stands for the set of all greatest-type divisors of x in S. The concept of greatest-type divisor was introduced by Hong and played a key role in his solution [6] of the Bourque-Ligh conjecture [1]. Korkee et al. [13] studied the invertibility of matrices in a more general matrix class, join matrices. Meanwhile, they provided a lattice-theoretic explanation for this solution of the Bourque-Ligh conjecture.

    This paper is organized as follows. In Section 2, we supply some preliminary results that are needed in the proofs of the main results of this paper. Then in Section 3, we show that for a certain class of gcd-closed (resp. lcm-closed) sets S, [Sa] is nonsingular. This provides evidences to Conjecture 1.1. In Section 4, we show that for a certain class of gcd-closed (resp. lcm-closed) sets S, (1Sa) is nonsingular.

    If f is an arithmetic function and m is any positive integer, we denote by 1/f the arithmetic function defined as follows:

    1f(m)={0if f(m)=0,1f(m)otherwise.

    First, we need a result which gives the formula for the determinant of the power LCM matrix on a gcd-closed set.

    Lemma 2.1. [8,Lemma 2.1] If S is gcd closed, then

    det[Sa]=nk=1x2akαa,k, (2.1)

    where

    αa,k=d|xkd (2.2)

    throughout this paper.

    Lemma 2.2. Let n be a positive integer. Then

    \mathop \sum \limits_{d|n} \left(\dfrac{1}{\xi_a}*\mu\right)(d) = n^{-a}.

    Proof. The result follows immediately from [7,Lemma 7] applied to f = 1/\xi_a .

    Lemma 2.3. If k = 1 , then \alpha_{a, k} = x_1^{-a} .

    Proof. Lemma 2.3 follows immediately from Lemma 2.2.

    We also need Hong's reduction formulas.

    Lemma 2.4. [8,Lemma 2.5] If S is gcd closed and k\ge 2 , then

    \begin{align} \alpha_{a,k} = x_k^{-a}+\overset{k-1}{\underset{t = 1}{\sum}}(-1)^t{\underset{1\le i_1 < \cdots < i_t\le k-1}{\sum}}(x_k,x_{i_1},...,x_{i_t})^{-a}. \end{align} (2.3)

    Lemma 2.5. If S is gcd closed and k\ge 3 , then

    \begin{align} \alpha_{a,k} = x_k^{-a}+\overset{k-2}{\underset{t = 1} {\sum}}(-1)^t{\underset{2\le i_1 < \cdots < i_t\le k-1} {\sum}}(x_k,x_{i_1},...,x_{i_t})^{-a}. \end{align} (2.4)

    Proof. Since S is gcd closed, (x_k, x_1, x_{i_1}, ..., x_{i_t}) = x_1 for 2\le i_1 < \cdots < i_t\le k-1 . By (2.3), one has

    \begin{align} \alpha_{a,k} = &x_k^{-a}-{\underset{1\le i\le k-1} {\sum}}{(x_k,x_i)}^{-a}+{\underset{1\le i < j\le k-1} {\sum}}{(x_k,x_i,x_j)}^{-a}\\ &+\cdots+ (-1)^t{\underset{1\le i_1 < \cdots < i_t\le k-1}{\sum}}{(x_k,x_{i_1},...,x_{i_t})}^{-a} +\cdots+(-1)^{k-1}{(x_k,x_1,...,x_{k-1})}^{-a}\\ = &-{(x_k,x_1)}^{-a}+{\underset{2\le i_1\le k-1}{\sum}}(x_k,x_1,x_{i_1})^{-a}+\cdots +(-1)^t{\underset{2\le i_1 < \cdots < i_{t-1}\le k-1}{\sum}}(x_k,x_1,x_{i_1},...,x_{i_{t-1}})^{-a}\\ &+\cdots+(-1)^{k-2}{\underset{2\le i_1 < \cdots < i_{k-3}\le k-1}{\sum}}{(x_k,x_1,x_{i_1},...,x_{i_{k-3}})}^{-a} +(-1)^{k-1}(x_k,x_1,...,x_{k-1})^{-a}\\ &+x_k^{-a}-{\underset{2\le i\le k-1}{\sum}} {(x_k,x_i)}^{-a}+{\underset{2\le i < j\le k-1}{\sum}} {(x_k,x_i,x_j)}^{-a}\\ &+\cdots+(-1)^t{\underset{2\le i_1 < \cdots < i_t\le k-1}{\sum}}{(x_k,x_{i_1},...,x_{i_t})}^{-a} +\cdots+(-1)^{k-2}{(x_k,x_2,...,x_{k-1})}^{-a}\\ = &(-1)\overset{k-2}{\underset{w = 0}{\sum}}\tbinom {k-2}{w}(-1)^wx_1^{-a}+x_k^{-a}+\overset{k-2} {\underset{t = 1}{\sum}}(-1)^t{\underset{2\le i_1 < \cdots < i_t\le k-1}{\sum}}{(x_k,x_{i_1},...,x_{i_t})}^{-a}\\ = &(-1)\times 0+x_k^{-a}+\overset{k-2} {\underset{t = 1}{\sum}}(-1)^t{\underset{2\le i_1 < \cdots < i_t\le k-1}{\sum}}{(x_k,x_{i_1},...,x_{i_t})}^{-a}\\ = &x_k^{-a}+\overset{k-2}{\underset{t = 1} {\sum}}(-1)^t{\underset{2\le i_1 < \cdots < i_t\le k-1}{\sum}}{(x_k,x_{i_1},...,x_{i_t})}^{-a} \end{align}

    as expected. This concludes the proof of Lemma 2.5.

    Lemma 2.6. If S is gcd closed, then

    \begin{align} \alpha_{a,k} = {\underset{J\subset G_S(x_k)}{\sum}}\dfrac{(-1)^{|J|}}{(\gcd(J\cup {x_k}))^a}. \end{align} (2.5)

    Proof. This follows immediately from [9,Theorem 1.2] applied to f = 1/\xi_a .

    Lemma 2.7. [3,Theorem 5.1] Let a and n be positive integers with a\ge 2 and n\le 8 . If S is gcd closed, then [S^a] is nonsingular.

    Lemma 2.8. [10,Lemma 2.2] The set S is lcm closed if and only if mS^{-1} is gcd closed.

    As usual, for any nonzero integer c and prime number p , \nu_p(c) is the greatest nonnegative integer \nu such that p^{\nu} divides c . Then \nu_p(c)\ge 1 if and only if p divides c .

    Theorem 3.1. Let S be gcd closed. If there are primes p_3, ..., p_n (not necessarily distinct) such that

    \begin{align} \nu_{p_k}(x_k) > \max\limits_{2\le i\le k-1}\{\nu_{p_k}(x_i)\}, \end{align} (3.1)

    then [S^a] is nonsingular.

    Proof. By Lemma 2.3, we have \alpha_{a, 1} = x_1^{-a} . By Lemma 2.4, one has

    \alpha_{a,2} = x_2^{-a}-(x_1,x_2)^{-a} = x_2^{-a}-x_1^{-a}

    since (x_2, x_1) = x_1 . Thus \alpha_{a, 1}\neq 0 and \alpha_{a, 2}\neq 0 . Now let k\ge 3 . For any 2\le i_1 < \cdots < i_t\le k-1\ (1\le t\le k-2) , it follows from (3.1) that

    \nu_{p_k}(x_k,x_{i_1},...,x_{i_t}) = \min\{\nu_{p_k}(x_k),\nu_{p_k}(x_{i_1}),...,\nu_{p_k}(x_{i_t})\} < \nu_{p_k}(x_k).

    Therefore,

    \begin{align} \frac{x_k}{(x_k,x_{i_1},...,x_{i_t})}\equiv0 \pmod {p_k}. \end{align} (3.2)

    Multiplying both sides of (2.4) by x_k^a , by (3.2), one has x_k^a\alpha_{a, k}\equiv1 \pmod {p_k} . Thus, \alpha_{a, k}\neq0\ (k\ge3) . So \alpha_{a, 1}\alpha_{a, 2}\cdots\alpha_{a, n}\neq0 . It can be deduced from Lemma 2.1 that [S^a] is nonsingular.

    Corollary 3.1. Let S be gcd closed. If x_i|x_j for all integers i, j with 1\le i < j\le n , then [S^a] is nonsingular.

    Proof. We show that there are primes p_3, ..., p_n satisfying (3.1). Let 3\le k\le n . For any prime p , we have \nu_p(x_k)\ge \nu_p(x_{k-1}) since x_{k-1}|x_k . Thus there exists a prime p_k such that \nu_{p_k}(x_k) > \nu_{p_k}(x_{k-1}) since x_{k-1} < x_k . Moreover, x_2|...|x_{k-1} implies that

    \nu_{p_k}(x_2)\le \nu_{p_k}(x_3)\le\cdots \le \nu_{p_k}(x_{k-1}).

    Hence \max_{2\le i\le k-1}\{\nu_{p_k}(x_i)\} = \nu_{p_k}(x_{k-1}) . Then (3.1) holds. The statement is true from Theorem 3.1.

    Corollary 3.2. Let S be gcd closed. If x_2/x_1, ..., x_n/x_1 are pairwise relatively prime, then [S^a] is nonsingular.

    Proof. Let 3\le k\le n . By the hypothesis, x_k/x_1 is prime to x_i/x_1 for 2\le i\le k-1 . Recall that x_k/x_1 > x_i/x_1\ (2\le i\le k-1) . Let p_k be any prime factor of x_k/x_1 . Then we know that p_k\nmid x_i/x_1\ (2\le i\le k-1) . Thus, \nu_{p_k}\left(x_i/x_1\right) = 0\ (2\le i\le k-1) . Therefore

    \max\limits_{2\le i\le k-1}\left\{\nu_{p_k}\left(\dfrac{x_i}{x_1}\right)\right\} = 0.

    So

    \max\limits_{2\le i\le k-1}\{\nu_{p_k}(x_i)\} = \nu_{p_k}(x_1).

    Since p_k|x_k/x_1 , one has

    \nu_{p_k}\left(\dfrac{x_k}{x_1}\right)\ge 1.

    This yields that

    \nu_{p_k}(x_k)\ge \nu_{p_k}(x_1)+1 > \nu_{p_k}(x_1).

    Then (3.1) holds and the proof of Corollary 3.2 is complete.

    Theorem 3.2. Let S be gcd closed and F_k = \{x\in S|x < x_k\ {\rm {and}}\ x|x_k\} . If

    \begin{align} x_k > {\rm {lcm}}(F_k), \end{align} (3.3)

    for k = 2, ..., n , then [S^a] is nonsingular.

    Proof. By (2.3), we have

    \begin{align} \alpha_{a,k} = \frac{1}{x_k^a}+\frac{t}{y^a}, \end{align} (3.4)

    where t is an integer and y = {\rm {lcm}}(F_k) . Then y|x_k . On the other hand, (3.3) yields y < x_k . Then there exists a prime factor p of x_k such that \nu_p(y) < \nu_p(x_k) . So we obtain that

    \begin{align} \nu_p\left(\dfrac{x_k}{y}\right) > 0. \end{align} (3.5)

    Multiplying by x_k^a on both sides of (3.4), by (3.5), we have \alpha_{a, k}x_k^a\equiv1 \pmod p . Therefore, \alpha_{a, k}\neq0 . So \alpha_{a, 1}\alpha_{a, 2}\cdots\alpha_{a, n}\neq0 . Thus the statement is true from Lemma 2.1.

    Corollary 3.3. Let S be gcd closed and H_k = \{x_i|1\le i\le k-1\} .If

    \begin{align} x_k > {\rm{lcm}}(H_k), \end{align} (3.6)

    for k = 2, ..., n , then [S^a] is nonsingular.

    Proof. Let F_k be defined as in Theorem 3.2. Since F_k\subseteq H_k , one has

    \begin{align} {\rm{lcm}}(H_k)\ge {\rm{lcm}}(F_k). \end{align} (3.7)

    Equations (3.6) and (3.7) yield (3.3). Then by Theorem 3.2, we know that [S^a] is nonsingular.

    Corollary 3.4. Let S be gcd closed.If there is at most one x_{i_k}\ (2\le i_k\le k-1) such that x_{i_k}|x_k , then [S^a] is nonsingular.

    Proof. Let F_k be defined as in Theorem 3.2. If there does not exist any integer i_k with 2\le i_k\le k-1 such that x_{i_k}|x_k , then we have {\rm{lcm}}(F_k) = x_1 . So (3.3) holds. It then follows from Theorem 3.2 that [S^a] is nonsingular. The statement is true for this case.

    If there exists exactly one x_{i_k}\ (2\le i_k\le k-1) such that x_{i_k}|x_k , then we have {\rm{lcm}}(F_k) = x_{i_k} . The condition x_{i_k} < x_k yields (3.3). The corollary follows immediately from Theorem 3.2 as expected. The statement is true for this case.

    Theorem 3.3. Let a and n be positive integers with a\ge 2 and n\ge 9 .If S is (n-8) -fold gcd closed, then [S^a] is nonsingular.

    Proof. First of all, any (n-8) -fold gcd-closed set S must satisfy that x_1|...|x_{n-7} and the set \{x_{n-7}, ..., x_n\} is gcd closed. Since G_S(x_k) = \{x_{k-1}\} for all 2\le k\le n-7 , by (2.5), we have \alpha _{a, k} = x_k^{-a}-x_{k-1}^{-a} . Hence \alpha _{a, k}\neq0 for all integers k with 1\le k\le n-7 . Now let n-6\le k \le n . One has x_{n-7}|x_k since \{x_{n-7}, ..., x_n\} is gcd closed. So G_S(x_k) equals the set of greatest-type divisors of x_k in the set S_k: = \{x_{n-7}, ..., x_k\} . Thus by (2.5) we have

    \begin{align} \alpha_{a,k} = {\underset{J\subset G_S(x_k)}\sum}\dfrac{(-1)^{|J|}}{(\gcd(J\cup\{x_k\}))^a} = {\underset{J\subset G_{S_k}(x_k)}\sum}\dfrac{(-1)^{|J|}}{(\gcd(J\cup\{x_k\}))^a}. \end{align} (3.8)

    Note that 2\le k-n+8\le 8 since n-6\le k \le n . So |S_k|\le 8 . Since S_k is gcd closed, it follows immediately from Lemma 2.7 that [S_k^a] is nonsingular. So \det[S_k^a]\neq0 . one can easily deduce from Lemma 2.1 that

    \begin{align} \alpha_{a,k} = {\underset{J\subset G_{S_k}(x_k)}\sum}\dfrac{(-1)^{|J|}}{(\gcd(J\cup\{x_k\}))^a}\neq0. \end{align} (3.9)

    By (3.8) and (3.9), we obtain that \alpha_{a, k}\neq0 for all integers k with n-6\le k \le n . So \alpha_{a, 1}\alpha_{a, 2}\cdots\alpha_{a, n}\neq0 . Then by Lemma 2.1, we derive that [S^a] is nonsingular as desired.

    This finishes the proof of Theorem 3.3.

    In the following theorem, we will show that [S^a] defined on a certain class of lcm-closed sets is nonsingular.

    Theorem 3.4. Let a and n be positive integers with a\ge 2 . Then each of the following is true:

    \rm (i). If n\le 8 and S is lcm closed, then [S^a] is nonsingular.

    \rm (ii). If n\ge 9 and S is (n-8) -fold lcm closed, then [S^a] is nonsingular.

    Proof. Let m be defined as in Definition 1.2. Since

    [x_i,x_j] = \dfrac{x_ix_j}{(x_i,x_j)} = \dfrac{1}{m}\cdot x_i\cdot\dfrac{m}{(x_i,x_j)}\cdot x_j = \dfrac{1}{m}\cdot x_i\cdot\left[\dfrac{m}{x_i}, \dfrac{m}{x_j}\right]\cdot x_j,

    we have

    [S^a] = \dfrac{1}{m^a}W[(mS^{-1})^a]W,

    where W = {\rm diag}(x_1^a, ..., x_n^a) . So

    \det[S^a] = \det[\left(mS^{-1}\right)^a] \prod\limits_{k = 1}^n\left(\dfrac{x_k^2}{m}\right)^a.

    Therefore, to show the desired result, it suffices to prove that \det[(mS^{-1})^a]\neq0 .

    First, we prove part (i). By Lemma 2.8, we know that the reciprocal set mS^{-1} is gcd closed. Lemma 2.7 tells us that [(mS^{-1})^a] is nonsingular. So \det[(mS^{-1})^a]\neq0 . This completes the proof of part (i).

    Consequently, we prove part (ii). Since S is (n-8) -fold lcm closed, by Definition 1.3, we know that mS^{-1} is (n-8) -fold gcd closed. Theorem 3.3 tells us that [(mS^{-1})^a] is nonsingular. So \det[(mS^{-1})^a]\neq0 . This finishes the proof of part (ii).

    In this section, we will show that (\dfrac{1}{S^a}) defined on a certain class of gcd-closed (resp. lcm-closed) sets is nonsingular.

    Theorem 4.1. Let a and n be positive integers with a\ge 2 . Then each of the following is true:

    \rm (i). If n\le 8 and S is gcd closed, then (\dfrac{1}{S^a}) is nonsingular.

    \rm (ii). If n\ge 9 and S is (n-8) -fold gcd closed, then (\dfrac{1}{S^a}) is nonsingular.

    Proof. Since

    \begin{align*} \dfrac{1}{(x_i,x_j)^a} = \dfrac{[x_i,x_j]^a}{x_i^ax_j^a}, \end{align*}

    we have

    \begin{align*} (\dfrac{1}{S^a})& = T[S^a]T, \end{align*}

    where

    \begin{align*} T = {\rm diag}\left(\dfrac{1}{x_1^a},...,\dfrac{1}{x_n^a}\right). \end{align*}

    So

    \begin{align} \det(\dfrac{1}{S^a}) = \det[S^a]\prod\limits_{k = 1}^nx_k^{-2a}. \end{align} (4.1)

    By Lemma 2.1 and (4.1), we get that

    \det(\dfrac{1}{S^a}) = \prod\limits_{k = 1}^n\alpha_{a,k}.

    So we only need to prove that \prod\limits_{k = 1}^n\alpha_{a, k}\neq0 .

    We prove part (i) as follows. Lemma 2.7 tells us that [S^a] is nonsingular. By Lemma 2.1, we know that \prod\limits_{k = 1}^n\alpha_{a, k}\neq0 . This concludes the proof of part (i).

    Now we prove part (ii). Since S is (n-8) -fold gcd closed, Theorem 3.3 tells us that [S^a] is nonsingular. By Lemma 2.1, we know that \prod\limits_{k = 1}^n\alpha_{a, k}\neq0 . This completes the proof of part (ii).

    Theorem 4.2. Let a and n be positive integers with a\ge 2 . Then each of the following is true:

    \rm (i). If n\le 8 and S is lcm closed, then (\dfrac{1}{S^a}) is nonsingular.

    \rm (ii). If n\ge 9 and S is (n-8) -fold lcm closed, then (\dfrac{1}{S^a}) is nonsingular.

    Proof. Let m be defined as in Definition 1.2. Since

    \dfrac{1}{(x_i,x_j)} = \dfrac{1}{m}\cdot\dfrac{m}{(x_i,x_j)} = \dfrac{1}{m} \cdot\left[\dfrac{m}{x_i},\dfrac{m}{x_j}\right],

    one deduces that

    \begin{align*} (\dfrac{1}{S^a}) = \dfrac{1}{m^a}[(mS^{-1})^a]. \end{align*}

    So

    \begin{align} \det(\dfrac{1}{S^a}) = \dfrac{1}{m^{na}}\det[(mS^{-1})^a]. \end{align} (4.2)

    Thus, it is sufficient to prove the truth of the claim by stating that \det[mS^{-1})^a]\neq0 . By the proof of Theorem 3.4, we know that the claim is true.

    This finishes the proof of Theorem 4.2.

    The authors would like to thank the anonymous referees for careful readings of the manuscript and helpful comments. The corresponding author K. Cheng was supported partially by China's Central Government Funds for Guiding Local Scientific and Technological Development (No. 2021ZYD0013) and the Research Initiation Fund for Young Teachers of China West Normal University (No. 412769).

    We declare that we have no conflict of interest.



    [1] K. Bourque, S. Ligh, On GCD and LCM matrices, Linear Algebra Appl., 174 (1992), 65–74. https://doi.org/10.1016/0024-3795(92)90042-9 doi: 10.1016/0024-3795(92)90042-9
    [2] K. Bourque, S. Ligh, Matrices associated with multiplicative functions, Linear Algebra Appl., 216 (1995), 267–275. https://doi.org/10.1016/0024-3795(93)00154-R doi: 10.1016/0024-3795(93)00154-R
    [3] W. Cao, On Hong's conjecture for power LCM matrices, Czech. Math. J., 57 (2007), 253–268. https://doi.org/10.1007/s10587-007-0059-3 doi: 10.1007/s10587-007-0059-3
    [4] S. Hong, LCM matrices on an r-fold gcd-closed set (Chinese), Journal of Sichuan University (Natural Science Edition), 33 (1996), 650–657.
    [5] S. Hong, On LCM matrices on gcd-closed sets, Se. Asian B. Math., 22 (1998), 381–384.
    [6] S. Hong, On the Bourque-Ligh conjecture of least common multiple matrices, J. Algebra, 218 (1999), 216–228. https://doi.org/10.1006/jabr.1998.7844 doi: 10.1006/jabr.1998.7844
    [7] S. Hong, Gcd-closed sets and determinants of matrices associated with arithmetical functions, Acta Arith., 101 (2002), 321–332. https://doi.org/10.4064/aa101-4-2 doi: 10.4064/aa101-4-2
    [8] S. Hong, Notes on power LCM matrices, Acta Arith., 111 (2004), 165–177. https://doi.org/10.4064/aa111-2-5 doi: 10.4064/aa111-2-5
    [9] S. Hong, Nonsingularity of matrices associated with classes of arithmetical functions, J. Algebra, 281 (2004), 1–14. https://doi.org/10.1016/j.jalgebra.2004.07.026 doi: 10.1016/j.jalgebra.2004.07.026
    [10] S. Hong, Nonsingularity of matrices associated with classes of arithmetical functions on lcm-closed sets, Linear Algebra Appl., 416 (2006), 124–134. https://doi.org/10.1016/j.laa.2005.10.009 doi: 10.1016/j.laa.2005.10.009
    [11] S. Hong, K. Shum, Q. Sun, On nonsingular power LCM matrices, Algebr. Colloq., 13 (2006), 689–704. https://doi.org/10.1142/S1005386706000642 doi: 10.1142/S1005386706000642
    [12] P. Haukkanen, J. Wang, J. Sillanp\ddot{a}\ddot{a}, On Smith's determinant, Linear Algebra Appl., 258 (1997), 251–269. https://doi.org/S0024-3795(96)00192-9
    [13] I. Korkee, M. Mattila, P. Haukkanen, A lattice-theoretic approach to the Bourque-Ligh conjecture, Linear Multilinear A., 67 (2019), 2471–2487. https://doi.org/10.1080/03081087.2018.1494695 doi: 10.1080/03081087.2018.1494695
    [14] M. Li, Notes on Hong's conjectures of real number power LCM matrices, J. Algebra, 315 (2007), 654–664. https://doi.org/10.1016/j.jalgebra.2007.05.005 doi: 10.1016/j.jalgebra.2007.05.005
    [15] H. Smith, On the value of a certain arithmetical determinant, P. Lond. Math. Soc., 7 (1875), 208–213. https://doi.org/10.1112/plms/s1-7.1.208 doi: 10.1112/plms/s1-7.1.208
    [16] J. Wan, S. Hu, Q. Tan, New results on nonsingular power LCM matrices, Electron. J. Linear Al., 27 (2014), 652–669. https://doi.org/10.13001/1081-3810.1927 doi: 10.13001/1081-3810.1927
    [17] G. Zhu, On the divisibility among power GCD and power LCM matrices on gcd-closed sets, Int. J. Number Theory, in press. https://doi.org/10.1142/S1793042122500701
    [18] G. Zhu, On a certain determinant for a U.F.D., Colloq. Math., unpublished work.
  • This article has been cited by:

    1. GUANGYAN ZHU, MAO LI, ON THE DIVISIBILITY AMONG POWER LCM MATRICES ON GCD-CLOSED SETS, 2023, 107, 0004-9727, 31, 10.1017/S0004972722000491
    2. Guangyan Zhu, Mao Li, Xiaofan Xu, New results on the divisibility of power GCD and power LCM matrices, 2022, 7, 2473-6988, 18239, 10.3934/math.20221003
  • Reader Comments
  • © 2022 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(1522) PDF downloads(75) Cited by(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog