For a simple connected graph G of order n, the distance signless Laplacian matrix is defined by DQ(G)=D(G)+Tr(G), where D(G) and Tr(G) is the distance matrix and the diagonal matrix of vertex transmission degrees, respectively. The zero divisor graph Γ(R) of a finite commutative ring R is a simple graph, whose vertex set is the set of non-zero zero divisors of R and two vertices v,w∈Γ(R) are edge connected whenever vw=wv=0. In this article, we find the DQ-eigenvalues of zero divisor graph of the ring Zn for general value n=pl11pl22, where p1<p2 are distinct prime numbers and l1,l2∈N. Further, we investigate the DQ-eigenvalues of zero divisor graphs of local rings and the rings whose associated zero divisor graphs are Hamiltonian. Also, we obtain the trace norm and the Wiener index of Γ(Zn) for some special values of n.
Citation: Bilal A. Rather, M. Aijaz, Fawad Ali, Nabil Mlaiki, Asad Ullah. On distance signless Laplacian eigenvalues of zero divisor graph of commutative rings[J]. AIMS Mathematics, 2022, 7(7): 12635-12649. doi: 10.3934/math.2022699
[1] | Fareeha Jamal, Muhammad Imran . Distance spectrum of some zero divisor graphs. AIMS Mathematics, 2024, 9(9): 23979-23996. doi: 10.3934/math.20241166 |
[2] | Ali N. A. Koam, Ali Ahmad, Azeem Haider, Moin A. Ansari . Computation of eccentric topological indices of zero-divisor graphs based on their edges. AIMS Mathematics, 2022, 7(7): 11509-11518. doi: 10.3934/math.2022641 |
[3] | Mohammad Hamidi, Irina Cristea . Hyperideal-based zero-divisor graph of the general hyperring Zn. AIMS Mathematics, 2024, 9(6): 15891-15910. doi: 10.3934/math.2024768 |
[4] | Bilal A. Rather, Fawad Ali, Nasim Ullah, Al-Sharef Mohammad, Anwarud Din, Sehra . Aα matrix of commuting graphs of non-abelian groups. AIMS Mathematics, 2022, 7(8): 15436-15452. doi: 10.3934/math.2022845 |
[5] | Jahfar T K, Chithra A V . Central vertex join and central edge join of two graphs. AIMS Mathematics, 2020, 5(6): 7214-7233. doi: 10.3934/math.2020461 |
[6] | Milica Anđelić, Saleem Khan, S. Pirzada . On graphs with a few distinct reciprocal distance Laplacian eigenvalues. AIMS Mathematics, 2023, 8(12): 29008-29016. doi: 10.3934/math.20231485 |
[7] | Zhen Lin . On the sum of the largest Aα-eigenvalues of graphs. AIMS Mathematics, 2022, 7(8): 15064-15074. doi: 10.3934/math.2022825 |
[8] | Hengbin Zhang . Automorphism group of the commuting graph of 2×2 matrix ring over Zps. AIMS Mathematics, 2021, 6(11): 12650-12659. doi: 10.3934/math.2021729 |
[9] | Huan Xu, Tao Yu, Fawaz E. Alsaadi, Madini Obad Alassafi, Guidong Yu, Jinde Cao . Some spectral sufficient conditions for a graph being pancyclic. AIMS Mathematics, 2020, 5(6): 5389-5401. doi: 10.3934/math.2020346 |
[10] | Igal Sason . Observations on graph invariants with the Lovász ϑ-function. AIMS Mathematics, 2024, 9(6): 15385-15468. doi: 10.3934/math.2024747 |
For a simple connected graph G of order n, the distance signless Laplacian matrix is defined by DQ(G)=D(G)+Tr(G), where D(G) and Tr(G) is the distance matrix and the diagonal matrix of vertex transmission degrees, respectively. The zero divisor graph Γ(R) of a finite commutative ring R is a simple graph, whose vertex set is the set of non-zero zero divisors of R and two vertices v,w∈Γ(R) are edge connected whenever vw=wv=0. In this article, we find the DQ-eigenvalues of zero divisor graph of the ring Zn for general value n=pl11pl22, where p1<p2 are distinct prime numbers and l1,l2∈N. Further, we investigate the DQ-eigenvalues of zero divisor graphs of local rings and the rings whose associated zero divisor graphs are Hamiltonian. Also, we obtain the trace norm and the Wiener index of Γ(Zn) for some special values of n.
Throughout this study, all graphs are simple, finite, and connected. A graph is symbolized by G=(V(G),E(G)), where V(G)={w1,w2,…,wn} represents its vertex set, whereas E(G) represents its edge set. Further, the number of elements in V(G) is the order n while the size m of G is the number of elements in E(G). We write u∼v if a vertex u is adjacent to a vertex v. The degree (valency) symbolized by dG(v) of a vertex v is the number of vertices incident on v. If every vertex of G has the same degree, it is referred to as a regular graph. The n×n matrix A=(αıȷ), where αıȷ=1 when ı is edge connected to ȷ, and 0 otherwise, is the adjacency matrix of G. Assume that Deg(G)=diag(δ1,δ2,…,δn) is the diagonal matrix, where δi=dG(vi), i=1,2,…,n is the vertex degrees of G. The real symmetric positive-definite matrix Q(G)=Deg(G)+A(G) is known as the signless Laplacian matrix, while its eigenvalue set including multiplicities is called the signless Laplacian spectrum of G. We denote the complete graph by Kn, for more notations and terminology, see [8].
The distance d(v,w) between two unique vertices w≠v is specified in G as the length of the smallest path connecting v and w. The diameter of G is defined as the greatest distance among any two of its vertices. The matrix D(G)=(d(v,w)) is said to be the distance matrix of G, while TrG(u1) is the transmission degree of u1 and it is equal to the total of the distances between u1 and all other vertices in G, i.e., TrG(u1)=∑w∈V(G)d(w,u1). If TrG(vi) (or simply Tri) is the transmission degree of vi∈V(G), the sequence {Tri},i=1,2,…,n is known as the transmission degree sequence of G.
Suppose Tr(G)=diag(Tr1,Tr2,…,Trn) is the diagonal matrix of vertex transmissions degree sequence of G. The authors of [6] presented the signless Laplacian for the distance matrix of G. The matrix DQ(G)=D(G)+Tr(G) is known as the distance signless Laplacian matrix of G. Also, DQ(G) is real symmetric positive-definite for n>2, so its eigenvalues are real and may be arranged say γ1≥γ2≥⋯≥γn, where γ1 is said to be the DQ-spectral radius. More about DQ-matrix can be seen in [7,15,16] and references therein.
For a commutative ring R with multiplicative identity 1(≠0), the zero divisor graphs of R, represented by Γ(R), is a simple, connected and undirected graph whose vertex set is the set of non-zero zero divisors of R, in which two vertices x1 and x2 are edge connected whenever x1x2=0. The zero divisor graphs including their adjacency and (distance) Laplacian eigenvalues have been studied in [5,9,10,16,18,20,21,25]. For eigenvalue analysis of other graphs defined on groups, see [2,3,22,23,24].
The rest of the manuscript is structured as follows. In Section 2, we start with some essential results and use them in proving our main problems. In Section 3, we deliberate the trace norms of Γ(Zn) of the DQ(G)-matrix for some special values of n.
Assume an n×n matrix
A=(A1,1A1,2⋯A1,mA2,1A2,2⋯A2,m⋮⋮⋱⋮Am,1Am,2⋯Am,m), |
such that, its columns and rows are partitioned according to a partition Π={π1,π2,…,πm} of I={1,2,…,n}. The quotient matrix Q (see [8]) of A is the matrix having m order, where (k,ℓ)-th entry is the average column sums (row sums) of Ak,ℓ. The partition Π is referred to as the equitable if every block Ak,ℓ has some constant column (row) sum, in such case, Q is known as the equitable quotient matrix. For equitable partitions, every eigenvalue of Q is also the eigenvalue of A.
Next, we have the definition of the joined union of graphs and state a result about DQ-spectrum of the joined union.
Definition 2.1. (Joined union) Assume G is an order n graph with vertex set V(G)={1,2,…,n} and Gi=Gi(Vi,Ei) are disjoint graphs having ni order, 1≤i≤n. The joined union G[G1,G2,…,Gn] of graphs, is obtained by considering graphs Gi,i=1,2,…,n and connect every vertex of Gk to each vertex of Gℓ, when k and ℓ are connected in G.
Theorem 2.2. [16] For a graph G with V(G)={u1,…,un}, and Gi is the ri-regular graphs having ni order whose adjacency eigenvalues are ri=λi1≥λi2≥…≥λini, whenever i=1,2,…,n. The DQ-spectrum of G[G1,…,Gn] contains the eigenvalues 2ni+n′i−ri−λik−4, for i=1,…,n and k=2,3,…,ni, when n′i=n∑k=1,k≠inkdG(ui,uk). The other n DQ-eigenvalues of G[G1,…,Gn] are the eigenvalues of following equitable quotient matrix:
Q=(4n1+n′1−2r1−4n2dG(u1,u2)…nndG(u1,un)n1dG(u2,u1)4n2+n′2−2r2−4…nndG(u2,un)⋮⋮⋱⋮n1dG(un,u1)n2dG(un,u2)…4nn+n′n−2rn−4). | (2.1) |
In general, it is very non trivial to determine the eigenvalues of any matrix. Here in algebraic theory of graphs, the eigenvalues of matrices corresponding to some special graphs like the complete bipartite graphs, the complete graphs are easily found. So, effort lies in transforming a graph by some operations into some nicely structures, so that the maximum eigenvalues of graph can be obtained. In [10], the authors showed that Γ(Zn) may be written as the joined union of graphs, where the components are either null graphs or cliques. The authors in [21] have found the structure of Γ(Zn) with n=pl11pl22.
Theorem 2.3. [21] For the zero divisor graph Γ(Zn) with n=pl11pl22, where p1<p2 are distinct primes and both l1=2s1, and l2=2s2 are positive even integers, where s1,s2≥1 are positive integers. The structure of Γ(Zn) is given as:
Γ(Zn)=Υn[¯Kϕ(pl1−11pl22),…,¯Kϕ(ps11pl22),…,¯Kϕ(pl22),¯Kϕ(pl11pl2−12),…,¯Kϕ(pl11ps22),…,¯Kϕ(pl11),¯Kϕ(pl1−11pl2−12),…,¯Kϕ(pl1−11ps22),…,¯Kϕ(pl1−11),…,¯Kϕ(ps11pl2−12),…,¯Kϕ(ps11ps2−12),Kϕ(ps11ps22),…,Kϕ(ps11),…,Kϕ(pl2−12),…,Kϕ(ps2−12),Kϕ(ps22),…,Kϕ(p2)], |
where Υn is referred to as the divisor graph since its vertices are defined as proper divisors of n and two vertices are connected if their product is a multiple of n.
From the Theorems 2.2 and 2.3, we see that out of n−1−ϕ(n) number of DQ-eigenvalues of Γ(Zn), n−1−t−ϕ(n) are positive integer, where t is the order of Υn, the other t DQ-eigenvalues of the graph Γ(Zn) are the eigenvalues of the equitable quotient matrix.
Next, we will illustrate the DQ-eigenvalues of the graph Γ(Zn) with n=pl11pl22, where p1<p2 are distinct primes and l1≤l2 are positive even integers. This generalize the results of [16] in a natural setting.
Theorem 2.4. For the graph Γ(Zpl11pl22), the DQ-spectrum of Γ(Zn) comprises of the eigenvalues
γi=2N+ϕ(pl11)pl2−12−pj1−3,fori=j=1,2,…,s1,…,l1−1,γl1=2N+ϕ(pl11)(pl2−12−1)−pl11−3,γi=2N+pl1−11ϕ(pl22)−pj2−3,forj=1,2,…,l2−1,andi=l1+1,…,l1+l2−1,γl1+l2=2N+(pl1−11−1)ϕ(pl22)−pl22−3,γi=2N−p1pj2−3,forj=1,2,…,l2,andi=l1+l2+1,…,l1+2l2,⋮γi=2N−ps11pj2−3,forj=1,2,…,s2−1,andi=l1+s2l2+1,…,l1+s2l2+s2−1,γi=2N−ps11pj2−1,forj=s2,…,l2,andi=l1+s2l2+s2,…,l1+(s2+1)l2⋮γi=2N−pl11pj2−1,forj=1,2,…,s2−1,andi=l1+l1l2+1,…,l1+l1l2γi=2N+ϕ(pl2−j2)−pl11pj2−1,forj=1,2,…,s2−1,andi=l1+l1l2+1,…,+s2−1γi=2N+ϕ(pl2−j2)−pl11pj2−1,forj=s2,…,l2,andi=l1+l1l2+s2,…,l1+l1l2+l2−1 |
with multiplicities ϕ(pl1−i1pl22)−1, ϕ(pl11pl2−j2)−1, ϕ(pl1−i1pl2−j2)−1, …, ϕ(ps11pl2−j2)−1, …,ϕ(pl2−j2)−1, respectively, where i=1,…,l1 and j=1,…,l2. The persisting DQ-eigenvalues of Γ(Zn) are the eigenvalues of matrix (2.1).
Proof. Let n=pl11pl22, where 2<p1<p2 are primes and 2≤l1=2s1≤2s2=l2, where s1 and s2 are positive even integers. Then by Theorem 2.3, we have
Γ(Zn)=Υn[¯Kϕ(pl1−11pl22),…,¯Kϕ(ps11pl22),…,¯Kϕ(pl22),¯Kϕ(pl11pl2−12),…,¯Kϕ(pl11ps22),…,¯Kϕ(pl11),¯Kϕ(pl1−11pl2−12),…,¯Kϕ(pl1−11ps22),…,¯Kϕ(pl1−11),…,¯Kϕ(ps11pl2−12),…,Kϕ(ps11ps22),…,Kϕ(ps11),…,Kϕ(pl2−12),…,Kϕ(ps22),…,Kϕ(p2)]. |
We shall now use Theorem 2.2, for calculating the DQ-eigenvalues of Γ(Zn). For that, we first need to know the values of n′i's. It is well established that zero divisor graphs of rings have a maximum diameter of three, so pi1∼pi2 if and only if n=i=j, otherwise pi1∼pk1pn2,k+i≥n and pj2∼pn1ph2,h+j≥n and finally pk1pn2∼pn1ph2,k≥1,h≥1. This means, d(pi1,pj2)=3,if1≤j,i≤n−1 in Υn, likewise distance between other vertices is at most 2. Now,
n′1=2(ϕ(pl1−21pl22)+⋯+ϕ(ps11pl22)+⋯+ϕ(pl22))+3(ϕ(pl11pl2−12)+⋯+(pl11ps22)+…+ϕ(pl11))+2(ϕ(pl1−11pl2−12)+⋯+ϕ(pl1−11ps22)+⋯+ϕ(pl1−11))+⋯+2(ϕ(ps11pl2−12)+⋯+ϕ(ps11ps22)+⋯+ϕ(ps11))+⋯+2(ϕ(pl2−12)+⋯+ϕ(ps22)+⋯+ϕ(p2))−ϕ(p1), |
where by definition of n′1, ϕ(pl1−11pl22) is removed and p1∼pl1−12pl22, so we subtract ϕ(p1). As ∑d|lϕ(d)=l, so order of Γ(Zn) is N=n−ϕ(n)−1=∑1,n≠d|nϕ(n). By applying Theorem 2.2, and using the number theory identities l∑i=1ϕ(pi1)=pl1−1 and ϕ(z1,z2)=ϕ(z1)ϕ(z2), if and only if (z1,z2)=1, we simplify the form of n′1 as:
n′1=2(N−ϕ(pl1−11pl22))+(ϕ(pl11pl2−12)+⋯+ϕ(pl11ps22)+⋯+ϕ(pl11))−ϕ(p1)=2(N−ϕ(pl1−11pl22))+ϕ(pl11)(ϕ(pl2−12)+⋯+ϕ(ps22)+⋯+ϕ(p2)+1)−ϕ(p2)=2(N−ϕ(pl1−11pl22))+ϕ(pl11)pl2−12−ϕ(p1). |
Now, by Theorem 2.2, the DQ-eigenvalues of Γ(Zn) are given as:
2n1+n′1−r1−λ1k−4=2ϕ(pl1−11pl22)+2(N−ϕ(pl1−11pl22)+ϕ(pl11)pl2−12−ϕ(p1)−0−0−4=2N+ϕ(pl11)pl2−12−ϕ(p1)−4. |
Thus, 2N+ϕ(pl11)pl2−12−ϕ(p1)−4 is the DQ-eigenvalue with multiplicity ϕ(pl1−11pl22)−1. Continuing in the same manner, other n′i's are given by:
n′i=2(N−ϕ(pl1−j1pl22))+ϕ(pl11)pl2−12−(pj1−1),fori=j=2,…,s1,…,l1−1,n′l1=2(N−ϕ(pl22))+ϕ(pl11)(pl2−12−1)−(pl11−1),n′i=2(N−ϕ(pl11pl2−j2))+ϕ(pl22)pl1−11−(pj2−1)fori=l1+1,…,l1+l2−1andj=1,…,s2,…,l2−1,n′l1+l2=2(N−ϕ(pl11))+ϕ(pl22)(pl1−11−1)−(pl22−1),n′i=2(N−ϕ(pl1−11pl2−j2))−(p1pj2−1),fori=l1+l2+1,…,l1+2l2andj=1,…,s2,…,l2,⋮n′i=2(N−ϕ(ps11pl2−j2))−(ps11pj2−1),fori=l1+s1l2+1,…,l1+s1l2+s2−1andj=1,…,s2−1,n′i=2N−ϕ(ps11pl2−j2)−(ps11pj2−1),fori=l1+s1l2+s2,…,l1+(s1+1)l2andj=s2,…,l2,⋮n′i=2(N−ϕ(pl2−j2))−(pl11pj2−1),fori=l1+l1l2+1,…,l1+l1l2+s2−1andj=1,…,s2−1,n′i=2N−ϕ(pl2−j2)−(pl11pj2−1),fori=l1+l1l2+s2,…,l1+l1l2+l2−1andj=s2,…,l2−1. |
Now, using the values of these n′i's and the Theorem 2.2, the other DQ-eigenvalues can be calculated as in statement. The rest DQ-eigenvalues of the graph Γ(Zn) are presented in the matrix (2.1).
In particular if l2=0, we have the sequel consequence of Theorem 2.4.
Corollary 2.5. For n=p2m1, m≥2 is a positive integer, the DQ-spectrum of Γ(Zn) contains of the eigenvalue 2N−pi1−3 having multiplicity ϕ(p2m−i1)−1, where i=1,…,m−1, and the eigenvalue N+(p2m−11−pi1)−2, for i=m,…,2m−1. The other DQ-eigenvalues of Γ(Zn) are the eigenvalues of the matrix (2.3).
Proof. The proper divisor set of n is {p1,p21,…,p2m−11}, we see that the vertex pi1 is connected to pj1 in Υn for any j≥2m−i where 1≤i≤2m−1 and j≠i. As n does not divide (pi1)2, for i=1,…,m−1, so
Gi={¯Kϕ(p2m−i1)fori=1,2,…,m−1,Kϕ(p2m−i1)fori=m,…,2m−1. | (2.2) |
From Eq (2.2), it follows that ni=ϕ(p2m−i1), where i=1,…,2m−1 and N=2m−1∑i=1ni. Also, by definition of n′i, we get
n′1=2n2+2n3+⋯+2n2m−2+n2m−1=22m−1∑i=2ni−n2m−1. |
Similarly, we obtain
n′i=22m−1∑j=1j≠inj−i∑j=1n2m−j,fori=1,…,m−1, |
and
n′i=2m−1∑j=1j≠inj+2m−1−i∑j=1nj,fori=m,…,2m−1. |
By Eq (2.2), we note that n1=¯Kϕ(p2m−11), so by Theorem 2.2, we see that
γ1=2n1+n′1−r1−λ1k−4=22m−1∑i=1ni−n2m−1−0−0−4=2N−ϕ(p1)−4=2N−p1−3 |
is the DQ-eigenvalue with multiplicity ϕ(p2m−11)−1. For i=2,3,…,m−1, proceeding as above with ni=¯Kϕ(p2m−i1), we get
γi=2N−i∑j=1ϕ(pj1)−4=2N−pi1−3, |
having multiplicities ϕ(p2m−i1)−1, where we use the property r∑i=1ϕ(pi1)=pr1−1 is used. Similarly, for i=m,…,2m−1, with Gi=Kϕ(p2m−i1),ri=ϕ(p2m−i1)−1 and λik=−1, the other DQ-eigenvalues are
γi=2ni+n′i−ri−λik−4=2ϕ(p2m−i1)+n′i−ϕ(p2m−i1)+1+1−4=2m−1∑j=1nj+2m−1−i∑j=1nj−2=N+(p2m−11−pi1)−2, |
having multiplicities ϕ(p2m−i1)−1. The rest DQ-eigenvalues are of the subsequent matrix:
(d12ϕ(p2m−21)⋯2ϕ(pm+11)2ϕ(pm1)2ϕ(pm−11)⋯2ϕ(p21)ϕ(p1)2ϕ(p2m−11)d2⋯2ϕ(pm+11)2ϕ(pm1)2ϕ(pm−11)⋯n2m−2ϕ(p1)⋮⋮⋱⋮⋮⋮⋱⋮⋮2ϕ(p2m−11)2ϕ(p2m−21)⋯dm−12ϕ(pm1)ϕ(pm−11)⋯ϕ(p21)ϕ(p1)2ϕ(p2m−11)2ϕ(p2m−21)⋯ϕ(pm+11)dmϕ(pm−11)⋯ϕ(p21)ϕ(p1)2ϕ(p2m−11)2ϕ(p2m−21)⋯ϕ(pm+11)ϕ(pm1)dm+1⋯ϕ(p21)ϕ(p1)⋮⋮⋱⋮⋮⋮⋱⋮⋮2ϕ(p2m−11)ϕ(p2m−21)⋯ϕ(pm+11)ϕ(pm1)ϕ(pm−11)⋯d2m−2ϕ(p1)2ϕ(p2m−11)ϕ(p2m−21)⋯ϕ(pm+11)ϕ(pm1)ϕ(pm−11)⋯ϕ(p21)d2m−1) | (2.3) |
where di={4ni+n′i−4,fori=1,2,…,m−1,4ni+n′i−2ri−4=2ni+n′i−2,fori=m,m+1,…,2m−1.
The topological indices are molecular descriptors used in the developments of quantitative structure activity relationships (QSARs), where molecular activities are related to the chemical structures of graphs. There are several well known topological indices, one such is the Wiener index introduced by Harry Wiener and has applications in chemical graph theory and computer networks (see [4,14]).
As sum of the eigenvalues of DQ(G) is equal to twice the Wiener index, that is, Trace(DQ(G))=2W, thus, we compute the Wiener index of Γ(Zn) using Theorem 2.4 and Corollary 2.5. First, we shall compute the Wiener index of Γ(Zp2m1).
From Corollary 2.5, the spectrum of Γ(Zp2m1) consists
{(2N−p1−3)[ϕ(p2m−11)−1],(2N−p21−3)[ϕ(p2m−21)−1],…,(2N−pm−11−3)[ϕ(pm+11)−1](2N−1−pm1)[ϕ(pm1)−1],(2N−pm+11−1)[ϕ(pm−11)−1],…,(2N−p2m−21−1)[ϕ(p21)−1],(N−2)[ϕ(p1)−1]} |
together with the eigenvalues of the matrix (2.3), where
N=2m−1∑i=1ϕ(p2m−i1)=∑1,n≠d|nϕ(d)=n−ϕ(n)−1=p2m1−ϕ(p2m1)−1=p2m−11−1. |
Now, the trace of the matrix (2.3) is d1+d2+⋯+dm−1+dm+dm+1⋯+d2m−2+d2m−1, where
di={4ni+n′i−4,fori=1,…,m−1,4ni+n′i−2ri−4=2ni+n′i−2,fori=m,m+1,…,2m−1. |
Also, d1=4n1+n′1−4=4ϕ(p2m−11)+22m−1∑i=2ϕ(p2m−i1)−ϕ(p1)−4=2N+2ϕ(p2m−11)−p1−3. Similarly, other di's are
di={2N+2ϕ(p2m−i1)−pi1−3fori=2,3,…,m−1,2N+ϕ(p2m−i1)−pi1−1fori=m,…,2m−1. |
Therefore, the trace of the matrix DQ(Γ(Zp2m1)) is given by:
Trace(DQ(Γ(Zp2m1)))=(2N−pi1−3)(ϕ(p2m−i1)−1)+(2N−pj1−1)(ϕ(p2m−j1)−1)+4N+2ϕ(p2m−i1)+ϕ(p2m−j1)−pi1−pj1−4. |
Thus the Wiener index of Γ(Zp2m1) is 12Trace(DQ(Γ(Zp2m1))).
Proceeding as above, the Wiener index of Γ(Zp2s11p2s22) can be found from its DQ-spectrum given in Theorem 2.4.
Similar to Theorem 2.4 and Corollary 2.5, the DQ-spectrum of Γ(Zpl11pl22) can be discussed, when both l1 and l2 are odd and when one of them is even and other is odd.
The following result demonstrates the DQ-eigenvalues of zero divisor graphs of some local rings. But before proceeding further, we need the following results.
Theorem 2.6. [7] The spectrum of DQ(Kn) is given below:
{(2n−2),(n−2)[n−1]} |
and that of DQ(Ka,b) is given by:
{5n−8±√9(a−b)2+4ab2,(2n−b−4)[a−1],(2n−a−4)[b−1]} |
where n=a+b.
A complete split graph, represented by CSω,n−ω, is a graph that consists of a clique on ω vertices while an independent set on the rest of n−ω vertices, so that any vertex of the clique is connected with each vertex of the independent set.
Theorem 2.7. [16] The DQ-eigenvalues of CSω,n−ω are given as:
{(n−2)[ω−1],(2n−ω−4)[n−ω−1],12(5n−2ω−6±√4ωn−6ω2+8ω−3n−2)}. |
Theorem 2.8. [Theorems 6 and 7, [1]] For a finite commutative ring R, if all the possible vertices of Γ(R) (or ¯Γ(R)) have the equal degrees, then either R≅F×F or Z(R)2={0}, for some finite field F.
Theorem 2.9. Suppose R is a finite commutative ring with unity 1(≠0). We have
(i) If |R|=p21, where p1 is any prime, then the DQ-spectrum of Γ(R) is either {(2p1−4),(p1−3)[p1−2]} or {(7p1−11),(3p1−7)[2p1−3]}.
(ii) If R is local having order p31, then the DQ-spectrum of Γ(R) is either {(2p21−3),(p21−3)[p21−2]} or
{(2p21−p1−5)[p21−p1−1],(p21−3)[p1−2],5p21−2p1−9±√9p41−20p31+2p21+12p1+12}. |
Proof. (ⅰ) If R is local, then either R≅Zp21 or R≅Zp1[x](x2) and in either case, Γ(R) is a complete graph whose order is p1−1. Thus, by Theorem 2.6, we get
DQ(Γ(R),x)=(x−2p1+4)(x−p1+3)p1−2. |
Therefore, spec(Γ(R)) is as desired. If R is reduced, then R≅Zp1×Zp1, and hence Γ(R) is complete bipartite. Thus, by Theorem 2.6, we get
DQ(Γ(R),x)=(x−(5p1−9±2(p1−1)))(x−3p1+7)2p1−4. |
Thus, the DQ-spectrum is as desired.
(ⅱ) If R is local and |R|=p31, then R is isomorphic to any of the subsequent: Fp1[x,y](x,y)2, Fp1[x](x3), Zp21[x](p1x,x2), or Zp21[x](p1x,x2−ˉsp1), where ˉs∈Zp1 is a non-square element. If R≅Fp1[x,y](x,y)2, then Z∗(R)={uy}∪{ux}∪{xu+yu′}, where u′,u∈Fp1∖{0}. Therefore, |Γ(Fp1[x,y](x,y)2)|=p21−1, and for every u,v∈Z∗(R), we have uv=0. Thus, Γ(R)=Kp21−1. Also, if R≅Zp21[x](p1x,x2), then Z∗(R)={xu}∪{p1u}, where u∈Zp1∖{0}, so Γ(R)≅Kp21−1. Thus in either case, when R≅Fp1[x,y](x,y)2 or Zp21[x](p1x,x2), then DQ-spectrum of Γ(R) is {(2p21−3),(p21−3)[p21−2]}. Next, if R≅Fp1[x](x3), then Z∗(R) can be partitioned into two subsets; Z1={ux2|u∈Fp1∖{0}} and Z2={ax+bx2|a∈Fp1∖{0},b∈Zp1}. Then Z1 induces a clique having p1−1 vertices while Z2 is an independent subset. Further, for every z1∈Z1 and z2∈Z2, we have z1z2=0. Finally, if R≅Zp21[x](p1x,x2−ˉsp1), where ˉs is a non-square element in Zp1, then the vertex set of Γ(Zp21[x](p1x,x2−ˉsp1)) can be expressed as disjoint union of the sets S1 and S2, where, S1={up1|u∈Zp1∖{0}} and S2={ux}∪{up1+u′x|u,u′∈Zp1∖{0}}. Then, ∀s1,s′1∈S1 and s2,s′2∈S2, we have s1s′1=0, s1s2=0 and s2s′2≠0. Thus, in each of these cases, Γ(R) is a complete split graph CSp1−1,p21−p1. Therefore, by Theorem 2.7, DQ-spectrum of Γ(R) is {(2p21−3),(p21−3)[p21−2]} or {(2p21−p1−5)[p21−p1−1],(p21−3)[p1−2],5p21−2p1−9±√9p41−20p31+2p21+12p1+12}.
Theorem 2.10. Suppose R is a finite commutative ring. If either Γ(R) (or ¯Γ(R)) is regular, then the DQ-spectrum of Γ(R) is either {(2|Z∗(R)|−4),(|Z∗(R)|−3)[|Z∗(R)|−2]} or {(7|Z∗(R)|−11),(3|Z∗(R)|−7)[2|Z∗(R)|−3]}.
Proof. If either Γ(R) (or ¯Γ(R)) is regular, then using Theorem 2.8, either Z(R)2=0 or there is a field F such that R≅F×F. If Z(R)2=0, then Γ(R)≅K|Z∗(R)|−1, then by Theorem 2.6, the DQ-spectrum of Γ(R) is obtained by replacing n by |Z∗(R)|−1. Further, if R≅F×F, then Γ(R)≅K|F−1|,|F−1|, and hence DQ-spectrum of Γ(R) is {(7|Z∗(R)|−11),(3|Z∗(R)|−7)[2|Z∗(R)|−3]}.
If a graph G contains a cycle that transverses each vertex, then G is said to be Hamiltonian.
Theorem 2.11. Suppose R≅R1×R2 is a finite commutative ring whose zero divisor graph is Hamiltonian, then the DQ-spectrum is given as:
{5|R|−18±√12(9|R|2−4|R|−32|R1||R2|+1),(2|R|+|Ri|−1)[|Rj|−1]}, |
where i,j∈{1,2}, and i≠j.
Proof. We prove that both R1 and R2 must be integral domains. If not, let Z1={0}×Z∗(R2) and Z2=(R1−Z(R1))×Z∗(R2). Then, Z2 is independent while there is z1∈Z1 and z2∈Z2 such that z1z2=0. Now, a Hamiltonian cycle in Γ(R) containing all vertices of Z2 and therefore containing a matching among Z1 and Z2. Since, Z2 is an independent set, it means |Z2|≤|Z1|. This means that |R1−Z(R1)|≤1, implying that the only unit in R1 is the identity element. Thus, R1≅ΠZk2 for some k∈N. Consider z′=(1,1,⋯,1,0)∈R1, then (z′,1)∈V(Γ(R1×R2)) is the only vertex which is connected to z″=(0,0,⋯,0,1,0), which is the contradiction with the fact that Γ(R) is Hamiltonian. As a result, both R1 and R2 are integral domains. Now, as R is finite, so Γ(R)≅K|R1|−1,|R2|−1. Therefore, by Theorem 2.6, the DQ-spectrum of Γ(R) is
{5|R|−18±√1/2(9|R|2−4|R|−32|R1||R2|+1),(2|R|+|Ri|−1)[|Rj|−1]}, |
where i,j∈{1,2}, and i≠j.
Suppose Mn(C) is the set of all n×n square matrices over the complex field C. For M∈Mn(C), the square roots of the eigenvalues of M∗M or MM∗ are called the singular values, where M∗ is the complex conjugate of M. As MM∗ is positive semi-definite, so the singular values of M are non-negative, denoted by s1(M)≥s2(M)≥⋯≥sn(M). The trace norm of M∈Mn(C) is specified as the sum of singular values, that is,
‖M‖n=s1(M)+s2(M)+⋯+sn(M), |
and the sum of the first k singular values is the Ky Fan k-norm, that is,
‖M‖k=s1(M)+s2(M)+⋯+sk(M). |
‖M‖1 is the largest singular value of M and is called the spectral norm. It is obvious that for a Hermitian matrix M, si(M)=|λi(M)|, and for a positive semi-definite matrix M, si(M)=λi(M), where the eigenvalues of M are λi(M), i=1,…,n.
The trace norm of the symmetric matrix DQ(G)−2W(G)nIn, is studied under the name distance signless Laplacian energy of G in the algebraic graph theory, where In is the identity matrix. For the symmetric matrix DQ(G)−2W(G)nIn, we have si(DQ(G)−2W(G)nIn)=|λi(DQ(G)−2W(G)nIn)| and the trace norm (distance signless Laplacian energy or DQ-energy) [11] of G is given below:
DSLE(G)=n∑i=1|γQi−2W(G)n|. | (3.1) |
Suppose σ is the largest positive integer such that γQσ≥2W(G)n and Sk(G)=b∑i=1γi is the Ky Fan k-norm (sum of k largest DQ-eigenvalues) of the matrix DQ(G). Then using n∑i=1γi=2W(G), Eq (3.1) can be written in terms of Ky Fan k-norm [11] as shown below
DSLE(G)=2(σ∑i=1γi−2σW(G)n)=2max1≤j≤n(j∑i=1γi(G)−2jW(G)n). |
For some latest works on DSLE(G), see [11,13].
Assume that G-uv is the connected graph attained from G by removing an edge uv. The next result states that the DQ-spectrum of G decreases upon edge deletion.
Lemma 3.1. [6] Suppose G is a simple graph whose order and size are n and m, respectively, where n≤m and G′=G−e is a connected graph attained from G by removing an edge. If γ1(G)≥γ2(G)≥⋯≥γn(G) and γ1(G′)≥γ2(G′)≥⋯≥γn(G′) are respectively the DQ-eigenvalues of G and G′. Then γi(G′)≥γi(G) satisfies for every 1≤i≤n.
Lemma 3.2. [16] Consider Γ(Zn) is the zero divisor graph. Then the following hold.
(i) The DQ-spectrum of Γ(Zp21) is {2p1−4,(p1−3)[p1−2]}.
(ii) The DQ-spectrum of Γ(Zp31) is
{(2p21−p1−5)[p21−p1−1],(p21−3)[p1−2],12(5p21−2p1−9±√9p41−20p31+2p21+12p1+1)}. |
The following result says that Γ(Zp21) has minimal DQ-energy among all the zero divisor graphs of order p1−1, where p1 is prime.
Theorem 3.3. Suppose Γ(R) is a zero divisor graph of ring R. Then
DSLE(Γ(R))≥2(k(p1−3)+p1−1−2kW(R)n), |
the equality holds iff n=p21, where p1 is prime.
Proof. As we know Γ(Zn) is complete if and only if n=p21. Thus by Lemma 3.1, γi(R)≥γi(Γ(Zp21)) ∀i=1,2,…,n. So,
Sk(Γ(R))≥Sk(Γ(Zp21))=k(p1−3)+p1−1. | (3.2) |
The equality holds if and only if n=p21. Let σ≥0 such that γσ≥2W(G))n. Then using Eq (3.2) and the definition of DQ-energy, we get
DSLE(Γ(R))=2(σ∑i=1γQi(Γ(R))−2σW(Γ(R))n)=2max1≤j≤n(j∑i=1γQi(Γ(R))−2jW(R)n)≥2max1≤j≤n(j∑i=1γQi(Γ(Zp21))−2jW(Γ(R))n)=2(k(p1−3)+p1−1−2(b−1)W(Γ(R))n), |
with equality as in (3.2).
Since 2W(Γ(Zp21))n=p1−2 and by Lemma 3.2, it is easy to see that σ=p1−1. The next consequence of Theorem 3.3 provides the DQ-energy of Γ(Zp21).
Corollary 3.4. The DQ-energy of Γ(Zp21) is
DSLE(Γ(Zp21))=p21−3p1+2. |
The next result states that among the class of zero divisor graphs of order N=p21−1 with independence and clique number is p21−p1 and p1−1, respectively, the graph Γ(Zp31) of Zp31 has the minimal trace norm.
Theorem 3.5. Assume that Γ(R) is a zero divisor graph of R with independence number p21−p1, with prime p1. Then
DSLE(Γ(R))≥2(√D+2p41−3p31−p21+4p1−4−2(p21−p1)W(Γ(R))N), |
where D=9p41−20p31+2p21+12p1+1, and equality occurs iff Γ(R)≅Γ(Zp31).
Proof. For n=p31, the only proper divisors of n are p1 and p21, so by the definition of zero divisor graph, Γ(Zp31)≅Kp1−1▽¯Kp21−p1, i.e., Γ(Zp31) is the complete split graph with independence number p21−p1. Thus by Lemma 3.1, γQi(Γ(R))≥γQi(Γ(Zp31)) for all i=1,2,…,(p21−2),(p21−1). Also, by Lemma 3.2, the DQ-spectrum of Γ(Zp31) is
{(2p21−p1−5)[p21−p1−1],(p21−3)p1−2,12(5p21−2p1−9±√9p41−20p31+2p21+12p1+1)}. |
Besides, it is easy to see that 2WN=2p41−2p31−3p21+p1+2p21−1. Let σ∈N such that γQσ≥2W(Γ(R))n. Since, 12(5p21−2p1−9+√9p41−20p31+2p21+12p1+1) is the DQ-spectral radius of Γ(Zp31) and is always greater than 2WN. Again, 2p21−p1−5≥2p41−2p31−3p21+p1+2p21−1 implies that p31−4p21+3≥0, which is true for p1>3. Now, if
12(5p21−2p1−9−√9p41−20p31+2p21+12p1+1)<2p41−2p31−3p21+p1+2p21−1, |
then we obtain (p21−1)√9p41−20p31+2p21+12p1+1−p41−2p31+8p21−5≥0, the inequality is true for p1>3. Similarly, the smallest DQ-eigenvalue p21−3 is always less than average of the eigenvalues. Thus, σ=1+p21−p1−1=p21−p1 and by definition of the DQ-energy, we have
DSLE(Γ(R))=2(σ∑i=1γQi(R)−2σW(R)N)≥2max1≤j≤n(j∑i=1γQi(Γ(Zp31))−2jW(R)N)=2(5p21−2p1−9+√D+(p21−p1−1)(2p21−p1−5)−2(p21−p1)W(Γ(R))N)=2(√D+2p41−3p31−p21+4p1−4−2(p21−p1)W(Γ(R))N), |
where D=9p41−20p31+2p21+12p1+1 and the equality holds if and only if Γ(R)≅Γ(Zp31).
The trace norm of Γ(Zp31) is obtained as a result of the previous theorem.
Corollary 3.6. The DQ-energy of Γ(Zp31) is given as:
DSLE(Γ(Zp21))=√9p41−20p31+2p21+12p1+1+2p41−7p31+p21+79+1p21−1. |
The parameter σ is very well studied [12] for different types of matrices associated with graphs. It is a very non trivial problem to characterize classes of graphs with particular σ and more interesting is to relate it with the parameters of a graph. There are rare graphs, where σ coincides with the independence number of graph. From Theorem 3.5, we see that Γ(Zp3) is one such family with σ same as the independence number.
The present article studies the distance signless Laplacian eigenvalues of the zero divisor graph Γ(Zpl11pl2) and the results are more general than in [16]. However, there are still gaps in the articel as all eigevalues of corresponding equitable quotient matrices cannot be found and in general the technique cannot be used for finding the distance signless Laplcain eigenvalues of Γ(Zn) as calculations become very hectic and majority of non integral distance signless Lpalacian eigenvalues of the corresponding matrix remains unknown. Some numerial methods may help in approximating the eigenvalue of equitable quotient matrix.
The author N. Mlaiki would like to thank Prince Sultan University for paying the publication fees for this work through TAS LAB.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
[1] | S. Akbari, A. Mohammadian, On zero-divisor graphs of finite rings, J. Algebra, 314 (2007), 168–184. |
[2] |
F. Ali, Y. Li, The connectivity and the spectral radius of commuting graphs on certain finite groups, Linear Multilinear A., 2019. https://doi.org/10.1142/S1793830918500350 doi: 10.1142/S1793830918500350
![]() |
[3] |
F. Ali, S. Fatima, W. Wang, On the power graph of certain of certain finite groups, Linear Multilinear A., 2020. https://doi.org/10.1080/03081087.2020.1856028 doi: 10.1080/03081087.2020.1856028
![]() |
[4] | F. Ali, Bilal A. Rather, A. Din, T. Saeed, A. Ullah, Power graphs of finite groups determined by Hosoya properties, Entropy, 24 (2022), 213. |
[5] |
D. F. Anderson, P. S. Livingston, The zero divisor graph of a commutative ring, J. Algebra, 217 (1999), 434–447. https://doi.org/10.1006/jabr.1998.7840 doi: 10.1006/jabr.1998.7840
![]() |
[6] | M. Aouchiche, P. Hansen, Two Laplacians for the distance matrix of a graph, Linear Algebra Appl., 439 (2013), 21–33. |
[7] | M. Aouchiche, P. Hansen, On the distance signless Laplacian of a graph, Linear Multilinear A., 64 (2016), 1113–1123. |
[8] | A. E. Brouwer, W. H. Haemers, Spectra of graphs, Springer, New York, 2010. |
[9] | I. Beck, Coloring of commutative rings, J. Algebra, 116 (1988), 208–226. |
[10] | S. Chattopadhyay, K. L. Patra, B. K. Sahoo, Laplacian eigenvalues of the zero divisor graph of the ring Zn, Linear Algebra Appl., 584 (2020), 267–286. |
[11] | K. C. Das, M. Aouchiche, P. Hansen, On (distance) Laplacian energy and (distance) signless Laplacian energy of graphs, Discrete Appl. Math., 243 (2018), 172–185. |
[12] | K. C. Das, S. A. Mojallal, Open problems on σ-invariant, Taiwanese J. Math., 23 (2019), 1041–1059. |
[13] |
R. C. Diaz, O. Rojo, Sharp upper bounds on the distance energies of a graph, Linear Algebra Appl., 545 (2018), 55–75. https://doi.org/10.11650/tjm/181104 doi: 10.11650/tjm/181104
![]() |
[14] |
A. A. Dobrynin, R. Entringer, I. Gutman, Wiener index of trees: Theory and applications, Acta Appl. Math., 66 (2001), 211–249. https://doi.org/10.1016/j.laa.2018.01.032 doi: 10.1016/j.laa.2018.01.032
![]() |
[15] |
H. Q. Liu, K. C. Das, Characterization of extremal graphs from distance signless Laplacian eigenvalues, Linear Algebra Appl., 500 (2016), 77–87. https://doi.org/10.1016/j.amc.2016.08.025 doi: 10.1016/j.amc.2016.08.025
![]() |
[16] |
S. Pirzada, Bilal A. Rather, M. Aijaz, T. A. Chisti, On distance signless Laplacian spectrum of graphs and spectrum of zero divisor graphs of Zn, Linear Multilinear A., 2020. http://doi.org/10.1080/03081087.2020.1838425 doi: 10.1080/03081087.2020.1838425
![]() |
[17] | S. Pirzada, Bilal A. Rather, T. A. Chishti, On distance Laplacian spectrum of zero divisor graphs of Zn, Carpathian Math. Publ., 13 (2021), 48–57. |
[18] |
S. Pirzada, Bilal A. Rather, T. A. Chishti, U. Samee, On normalized Laplacian spectrum of zero divisor graphs of commutative ring Zn, Electronic J. Graph Theory Appl., 9 (2021), 331–345. http://doi.org/10.5614/ejgta.2021.9.2.7 doi: 10.5614/ejgta.2021.9.2.7
![]() |
[19] | S. Pirzada, B. A. Rather, R. U. Shaban, Merajuddin, On graphs with minimal distance signless Laplacian energy, Acta Univ. Sapientae Math., 13 (2021), 450–467. |
[20] |
S. Pirzada, B. A. Rather, R. U. Shaban, Merajuddin, On signless Laplacian spectrum of zero divisor graphs of the ring Zn, Korean J. Math., 29 (2021), 13–24. https://doi.org/10.2478/ausm-2021-0028 doi: 10.2478/ausm-2021-0028
![]() |
[21] |
B. A. Rather, S. Pirzada, T. A. Naikoo, Y. Shang, On Laplacian eigenvalues of the zero divisor graph associated to the ring of integers modulo n, Mathematics, 9 (2021), 482. http://doi.org/10.3390/math9050482 doi: 10.3390/math9050482
![]() |
[22] | B. A. Rather, S. Pirzada, M. Imran, T. A. Chishti, On Randić Eigenvalues of zero divisor graphs of Zn, Commun. Comb. Optim., 2021, 1–12, http://doi.org/110.22049/CCO.2021.27202.1212 |
[23] |
B. A. Rather, S. Pirzada, T. A. Chishti, A. M. A. Alghamdi, On normalized Laplacian eigenvalues of power graphs associated to finite cyclic groups, Discrete Math. Algorithms Appl., 2022, 2250070. https://doi.org/10.1142/S1793830922500707. doi: 10.1142/S1793830922500707
![]() |
[24] |
B. A. Rather, S. Pirzada, T. A. Naikoo, On distance signless Laplacian spectra of power graphs of the integer modulo group, Art Discrete Appl. Math., 2022. https://doi.org/10.26493/2590-9770.1393.2be doi: 10.26493/2590-9770.1393.2be
![]() |
[25] |
M. Young, Adjacency matrices of zero divisor graphs of integer modulo n, Involve, 8 (2015), 753–761. https://doi.org/10.26493/1855-3974.93.e75 doi: 10.26493/1855-3974.93.e75
![]() |
1. | Bilal A. Rather, Fawad Ali, Nasim Ullah, Al-Sharef Mohammad, Anwarud Din, , Aα matrix of commuting graphs of non-abelian groups, 2022, 7, 2473-6988, 15436, 10.3934/math.2022845 | |
2. | Bilal Ahmad Rather, Fawad Ali, Suliman Alsaeed, Muhammad Naeem, Hosoya Polynomials of Power Graphs of Certain Finite Groups, 2022, 27, 1420-3049, 6081, 10.3390/molecules27186081 | |
3. | Fawad Ali, Bilal A. Rather, Nahid Fatima, Muhammad Sarfraz, Asad Ullah, Khalid Abdulkhaliq M. Alharbi, Rahim Dad, On the Topological Indices of Commuting Graphs for Finite Non-Abelian Groups, 2022, 14, 2073-8994, 1266, 10.3390/sym14061266 | |
4. | Bilal Ahmad Rather, Fawad Ali, Asad Ullah, Nahid Fatima, Rahim Dad, Aγ Eigenvalues of Zero Divisor Graph of Integer Modulo and Von Neumann Regular Rings, 2022, 14, 2073-8994, 1710, 10.3390/sym14081710 | |
5. | Clement Johnson, Ravi Sankar, Graph energy and topological descriptors of zero divisor graph associated with commutative ring, 2023, 1598-5865, 10.1007/s12190-023-01837-z | |
6. | Fareeha Jamal, Muhammad Imran, Distance spectrum of some zero divisor graphs, 2024, 9, 2473-6988, 23979, 10.3934/math.20241166 | |
7. | R. Sarathy, J. Ravi Sankar, Applications on color (distance) signless laplacian energy of annihilator monic prime graph of commutative rings, 2024, 15, 20904479, 102469, 10.1016/j.asej.2023.102469 | |
8. | Nadeem ur Rehman, Mohd Nazim, Exploring normalized distance Laplacian eigenvalues of the zero-divisor graph of ring Zn, 2023, 0009-725X, 10.1007/s12215-023-00927-y |