Skip to main content
Heliyon logoLink to Heliyon
. 2023 Sep 7;9(9):e19820. doi: 10.1016/j.heliyon.2023.e19820

Symmetries and symmetry-breaking in arithmetic graphs

Aqsa Shah 1, Imran Javaid 1,, Shahid Ur Rehman 1
PMCID: PMC10559213  PMID: 37809770

Abstract

In this paper, we study symmetries and symmetry-breaking of the arithmetic graph of a composite number m, denoted by Am. We first study some properties such as the distance between vertices, the degree of a vertex and the number of twin classes in the arithmetic graphs. We describe symmetries of Am and prove that the automorphism group of Am is isomorphic to the symmetric group Sn of n elements, for m=i=1npi. For symmetry-breaking, we study the concept of the fixing number of the arithmetic graphs and give exact formulae of the fixing number for the arithmetic graphs Am for m=i=1npiri under different conditions on ri.

MSC: 05C30, 05C12, 05A18

Keywords: Arithmetic graph, Automorphism, Symmetry-breaking, Twins, Fixing number

1. Introduction

Symmetries of graphs play central role in describing their structures and manipulating objects on graphs using their topologies. Distinguishability and indistinguishability of the vertices of graphs remain important problem for study which provide insights about the vertices and the structure of a graph. For this reason, identification of symmetries, breaking symmetries and distinguishing the vertices of graphs have always been interesting problems and have been well-studied in literature [1], [2], [3]. These concepts find applications in pattern recognition [4], strategies in mastermind games [5], problem of programming a robot for handling of objects [6] and coin weighing problem [7], [8]. Breaking the symmetries of a graph reduces the computational time and unnecessary computations by removing the symmetry-induced redundancies. Formally, each symmetry of a graph is a permutation of its vertex set which preserves the adjacency and non-adjacency relation of vertices in the structure, usually called the automorphism of a graph. To distinguish the vertices of a graph, symmetry, distance and degree related parameters in graphs using their topologies find more attention and these parameters yield interesting insights about graphs [9], [10], [11], [12], [13], [14]. A set sufficient to identify all symmetries of a graph was introduced by Boutin in [15] under the name determining set. If images of all the vertices of a determining set of a graph are known, then it is sufficient to identify the graph's automorphism group. On the other hand, fixing the vertices, i.e. imposing condition on the vertices that they will be mapped onto themselves, destroys symmetries. The minimum number of vertices that are needed to be fixed to destroy all the symmetries of a graph is called the fixing number and such a set of vertices is called a minimum fixing set. Boutin in [15] and Erwin and Harary in [6] independently introduced the terms of the determining number and the fixing number of graphs respectively. Boutin established in [15] that fixing the determining set will yield only identity automorphism which establishes the equivalence of the determining number and the fixing number. Hence, identifying symmetries and destroying symmetries are equivalent problems. The fixing numbers of different families of graphs, including Cartesian products [1], Cayley graphs and Frucht graphs [3] and complete graphs, paths, cycles [6] and as well as automorphism related parameters have been studied in [16], [17], [18], [19], [20], [21], [22]. The concept of symmetry-breaking was introduced by Albertson and Collins [23] using the concept of the distinguishing number of graphs where graphs vertices are labelled using t different labels so that no graph automorphism preserves labels. Minimum such t is called the distinguishing number and in [24], it is shown that the distinguishing number is strongly related to the idea of determining sets.

To distinguish the vertices of a given graph, Harary and Melter [25] and Slater [26] independently introduced the notions of resolving sets and the metric dimension, the minimum number of vertices that enable all other vertices to be uniquely distinguished by their distances from a subset of vertices in the graph. Harary and Melter [25] used the terminology of locating set and the location number whereas Slater [26] used the terms of resolving set and the metric dimension which we will be using wherever needed in this paper. In 2006, it was noticed by Erwin and Harary and also by Boutin that if all the vertices of resolving sets are mapped onto themselves, only trivial automorphism exists and that the minimum cardinality of a resolving set yields an upper bound for the fixing number of graphs. This bound is attained for graphs like paths, cycles, complete graphs and complete bipartite graphs and the metric dimension of trees and wheels is greater than the fixing numbers of these families of graphs. In 2006, Boutin posed the question “Can the difference between a graph's determining number and its smallest distance determining sets size be arbitrarily large?”. The question was addressed in [27] and [28]. Caceres et al. [27] in 2010, studied the difference of the metric dimension and the fixing number of cartesian products of graphs and trees. They showed that there are trees for which this difference is arbitrarily large. Garijo et al. [28], also studied the maximum value of the difference between the fixing number and the metric dimension of a graph. Recently, several authors studied both of these important parameters, the fixing number and the metric dimension of some graphs. In [29], the lower and upper bounds of fixing sets of the twin graphs are studied with some special classes of graphs. Das [30] investigated the fixing number of generalised and double generalised Petersen graphs. The fixing number and the metric dimension of the zero-divisor graph Γ(i=1nZ2) and the co-prime graphs with their properties are studied in [31], [32]. Due to equivalence of the determining number and the fixing number, we will use terms fixing set and the fixing number throughout the paper.

Now, we recall some definitions of graph theory which are necessary for this article. Graphs are used to model relations between elements of sets. Elements of sets are represented by vertices and relations between elements are represented by a set of edges (or links). Let G=(V(G),E(G)) be a simple graph where V(G) is a set of elements called vertices and E(G) is a set of relations called edges. The cardinality of V(G) and the cardinality of E(G) are called the order and the size of G, respectively. If there exists a set of edges from vertex x to vertex y, then we will say that there is a path between x and y. We say that G is connected if there is a path between each of its two vertices. The number of edges in the shortest xy path for a connected graph G is the distance between two vertices x and y, denoted by d(x,y).

An automorphism of a graph G for a graph G is a bijective mapping θ:V(G)V(G) such that θ(x)θ(y)E(G) if and only if xyE(G). The automorphism group of G, formed under the operation of composition of mappings is the set of all automorphisms defined on a graph G and is denoted by Aut(G). The stabilizer of a set F is a subset of Aut(G) and is defined as Stab(F)={θAut(G)|y=θ(y),yF}. The orbit of a vertex x, denoted by O(x), is the set defined as {xV(G)|θ(x)=yforsomeθAut(G)}. If two vertices x and y belongs to the same orbit, then they are similar vertices. A set FV(G) is a fixing set of G if the only automorphism that fixes every vertex in F is the identity automorphism and the fixing number of G, denoted by fix(G), is defined as the minimum cardinality of a fixing set. A vertex is fixed in a graph G, if it is fixed by every automorphism of G.

Numerous authors have investigated the connections between number theory and graph theory; for example, see [33], [34], [35]. We notice that graphs associated to different numbers display similar characteristics. Vasumathi et al. [36], introduced the idea of an arithmetic graph Am, studied some of its properties and also discovered an efficient method for creating an arithmetic graph using the given domination parameters. The arithmetic graphs containing the set of all divisors of m (excluding 1), where mN as its vertex set and m=p1r1p2r2...pnrn, for n distinct primes and ri1 for 1in. If two different divisors x,y of m have the same prime factors, such as (x=p1p2andy=p12p22), then they are said to have the same parity. Then xy if x,y are of different parity and gcd(x,y)=pi for i, 1in. Also, for any vertex x=i=1npiαi with at least one αi0, pi is termed as the primary factor whereas piαi, αi2 is termed as the secondary factor of x. The vertex set of the arithmetic graph Am for a composite number m=i=1npiri is denoted by V(Am) and the cardinality of the vertex set of the arithmetic graphs Am is given as |V(Am)|=i=1n(ri+1)1. The arithmetic graphs are the twin-free graphs when m=i=1npi and m=i=1npiri with 1ri2. Various authors examined the arithmetic graphs for different graph parameters [37], [38], [39]. In [40], Rao et al. investigated the split domination for arithmetic graphs. Moreover, the annihilator domination has been investigated for the arithmetic graphs in [41]. Rehman et al. [42], studied the arithmetic graphs and computed the results for the metric dimension of the arithmetic graphs Am.

Motivated by the close relationship between the fixing number and the metric dimension, we study symmetries and symmetry breaking of Am in this paper. For the purpose of this paper, we introduce the term, the skeleton of a vertex xV(Am), denoted by Sx. The skeleton Sx is the set of all those primes which are used in the prime factorization of x for Am. We partition the vertex set of Am for m=i=1npiri, into n classes Xl, (1ln), where Xl={xV(Am):|Sx|=l}. For example, if n=3 and ri=1 for 1i3 then X1={p1,p2,p3}, X2={p1p2,p2p3,p1p3} and X3={p1p2p3}. For our purpose, we write the canonical representation of m=i=1npiri such that r1r2rn. For example, if m=600 then p1=3,p2=5,p3=2 with r1=1,r2=2,r3=3.

The paper is organized as follows. In section 2, we study the properties of the arithmetic graphs in which we discuss the distance between any two vertices, the degrees of vertices and give the formula for twin classes in arithmetic graphs. Section 3 is based on the identification of symmetries of the arithmetic graphs and their related properties. We describe symmetries of the arithmetic graph Am and prove that Aut(Am) is isomorphic to Sn for m=i=1npi. We also give the automorphisms of Am for m=i=1npiri with ri1 and n2 by using permutations of {p1,p2,,pn}. In section 4, the connections between the fixing numbers and twin sets of Am are studied. We find the exact fixing number of the arithmetic graphs where m=i=1npiri with ri1. In the last section, concluding remarks are given.

2. Properties of the arithmetic graphs

In this section, we describe the properties of the arithmetic graphs such as the distance between vertices, the degree of a vertex and the number of twin classes in arithmetic graphs. We begin our study by giving results on the distance between the vertices of Am.

Lemma 2.1

Let m=i=12piri with ri1 then,

(i) For each x,yXi , 1i2 then d(x,y)=2 ,

(ii) For each xX2 and yX1 , then d(x,y)=1 or d(x,y)=3 .

Proof

(i) Suppose x,yXi, for 1i2 such that xy then we have the following cases.

Case 1. Let x,yX1 and SxSy then xp1p2y gives that d(x,y)=2.

Case 2. Let x,yX2 and Sx=Sy then xpiy, for any 1i2, gives that d(x,y)=2.

(ii) If xX2 and yX1 such that xy then d(x,y)=1. Now suppose that x=p1α1p2α2X2 for 1αi2 and at least one of αi1 for 1i2 and y=piβiX1, 1i2 and for some βi2 such that xy then xpip1p2y gives that d(x,y)=3. □

In the following lemma, we investigate the distance between any two vertices in Am for m=i=1npiri with n3 and ri1.

Lemma 2.2

Let m=i=1npiri with ri1 and n3 then,

(i) For each xXn and yV(Am)X1 , d(x,y)=2 .

(ii) For xy and x,yi=1n1Xi , 1d(x,y)2 .

(iii) For xXn and yX1 , d(x,y)=1 or d(x,y)=3 .

Proof

(i) Since Sx=Sm so SySx and |Sy|2 gives that gcd(x,y)pi which yield that xy. Let pjSy then ypj and xpi for each 1in gives that d(x,y)=2.

(ii) Now to prove that 1d(x,y)2 for xyi=1n1Xi, we study the following cases:

Case 1. Let x,yX1, then xy and we have the following cases:

Subcase 1. If SxSy and Sx={pi}, Sy={pj} for some 1ijn, then xpipjy which gives that d(x,y)=2.

Subcase 2. If Sx=Sy={pi} for some i, then xpipjy which gives that d(x,y)=2.

Case 2. Let xX1 and yi=2n1Xi such that xy, then we have the following cases:

Subcase 1. If x=pi for some 1in and xy so for every pjSy, xpipjy which yields that d(x,y)=2.

Subcase 2. If x=piα where 2αri for some i and xy so we have following possibilities:

(i) Let piSy. Since yi=2n1Xi so there exist at least one pjSm such that pjSy then xpipjy which gives that d(x,y)=2.

(ii) Let piSy and pjSy such that xpipjy which gives that d(x,y)=2.

Case 3. Let x,yi=2n1Xi and xy then we have following cases:

Subcase 1. Let SxSy=, piSx and pjSy then xpipjy which gives that d(x,y)=2.

Subcase 2. Let SxSy= and piSxSy then xpipjy which gives that d(x,y)=2.

Combining all above cases, we conclude that 1d(x,y)2.

(iii) For xXn and yX1, SySx. If x=p1p2...pn and y=piri for some 1in;ri1 then d(x,y)=1. For x=i=1npiαi for at least one αi2 and y=piβi for some 1in;βi2. Then, gcd(x,y)pi which yields that xy. Therefore, xpip1p2...pny which gives that d(x,y)=3. □

For m=i=12piri with all ri1, 1ij2 and 2αi,αjri, we notice that deg(pi)=deg(pj)=rirj, deg(pipj)=ri+rj, deg(piαi)=rj, deg(pjαj)=ri, deg(piαipj)=1+rj, deg(pipjαj)=ri+1 and deg(piαipjαj)=2. For the degrees of vertices of the arithmetic graph Am when m=i=1npiri with all ri1 and n3, we introduce the set X={1,2,,n} as the indexing set of m. For P={pi|iX} and AX, let mA=iApiri. Moreover, mA=0 if A=. For a non-identity divisor x=i=1npiαi of m, the collection Ix={iX|αi=1}, Jx={iX|αi2} and Kx={iX|αi=0} gives a partition of X. Note that AmKx is the induced subgraph of Am on mKx vertices and |V(AmKx)|=iKx(ri+1)1. In the next result, we give the formulae for the degree of x in Am.

Theorem 2.3

Let Am be the arithmetic graph of m=i=1npiri with n3 and ri1 for each i. For a vertex x=i=1npiαi for some 0αiri , we have

deg(x)={rj(iKx(ri+1)1)ifx=pjiKx(ri+1)1ifx=pjαj,αj2(iIxri+|Jx|)iKx(ri+1)ifxX1.

Proof

(i) Suppose x=pj, then Ix={j}, Jx= and Kx=X{j}. By definition of Am, x is not adjacent to any vertex in X1 and for each rjαj1, x is adjacent to every vertex of the form pjαjy for each yV(AmKx) which yields that there are rj|V(AmKx)| number of such vertices in V(Am)X1. Hence, deg(x)=rj|V(AmKx)|=rj(iKx(ri+1)1).

(ii) Suppose x=pjαj with αj2, then Ix=, Jx={j} and Kx=X{j}. By definition of Am, x is not adjacent to any vertex in X1 and x is adjacent to every vertex of the form pjy for each yV(AmKx) which yields that there are |V(AmKx)| number of such vertices in V(Am)X1. Hence, deg(x)=|V(AmKx)|=iKx(ri+1)1.

(iii) Now suppose x has at least two primary factors and x=i=1npiαi with 0αiri. For each iIx, the number of vertices adjacent with x in X1 with factor pi is ri and for each jJx the number of vertices adjacent with x in X1 with factor pj is one so the total number of vertices adjacent with x in X1 is iIxri+|Jx|. Also for each iIx and 1αiri, x is adjacent to every vertex of the form pjαjy for each yV(AmKx) and there are iIxri|V(AmKx)| number of such vertices in i=2nXi. Now for each jJx, x is adjacent to every vertex of the form pjy for each yV(AmKx) and there are |Jx||V(AmKx)| such vertices in i=2nXi. Hence, the total number of vertices adjacent with x in i=2nXi is iIxri|V(AmKx)|+|Jx||V(AmKx)| and the degree of x can be obtained as iIxri|V(AmKx)|+|Jx||V(AmKx)|+iIxri+|J|=(iIxri+|Jx|)(|V(AmKx)|+1). □

In the following corollary, we give the degrees of vertices in Am when m=i=1npi.

Corollary 2.4

Form=i=1npiwithn2. Then, we have

(i)deg(pi)=2n11,

(ii)deg(pi1pi2pi3...pik)=k(2)nk, with2knwhereij{1,2,...,n}and1jk.

Proof

(i) By definition of the arithmetic graph, pi is adjacent with vertices which have pi as a factor. Notice that pi has (n11) neighbours of the form pipi1, (n12) neighbours of the form pipi1pi2 and (n1nk) neighbours of the form j=1kpij has (n11)+(n12)+...+(n1n1) neighbours. It is a established fact that (n11)+(n12)+...+(n1n1)=2n11. Hence, deg(pi)=2n11 for 1in.

(ii) By definition of the arithmetic graph, j=1kpij can not be adjacent with a vertex x if it has two or more primes in common with j=1kpij. Therefore, j=1kpij has neighbours which have only one prime factor pij where 1jk and other factors are from remaining nk primes. j=1kpij is adjacent with vertices of the form pij where 1jk. j=1kpij has k(nk) neighbours of the form pijpl1 where pl1pij for any 1jk, k(nk2) neighbours of the form pijpl1pl2 where pltpij for any 1jk and t=1,2 and similarly k(nkq) neighbours of the form pijpl1pl2...plq where pltpij for any 1jk and 1tq. So deg(pi1pi2pi3...pik)=k((nk0)+(nk1)+(nk2)+...+(nknk))=k(2)nk which completes the proof. □

From Corollary 2.4(ii), it can be seen that all the vertices of Xk for 2kn have the same degree. It is important to notice that there may exist distinct vertices x,yV(Am) such that deg(x)=deg(y) and xXi and yXj where ij. For example, when m=i=13pi, deg(p1)=deg(p1p2p3).

A vertex x of G has the open neighbourhood if N(x)={yV(G):xyE(G)} and the closed neighbourhood if N[x]=N(x){x}. If N[x]=N[y], then two vertices x,y are true twins otherwise, they are false twins if N(x)=N(y). If every pair of vertices in a set of vertices T are twins, the set is referred to as a twin-set. The set of all twins of x is denoted by T(x) and the count of twin classes of graph G is denoted by TG. Note that T(x)={yV(Am):N[x]=N[y]orN(x)=N(y)}. A twin class is known as a trivial twin class if T(x)={x} whereas a non-trivial twin class has more than one vertex in it. A twin-free graph is a graph in which T(x)={x} for all xV(Am). For m=i=1npiri with n3 and ri1 for each i, T(j=1kpij)={j=1kpij} with 1kn and if rij=2 then T(j=1kpijαij)={j=1kpijαij} with 1αij2. If rij3 then |T(j=1kpijαij)|2 with αijrij and at least one αij3. Notice that for m=i=1npi and m=i=1npiri with n3 and 1ri2 are the twin-free arithmetic graphs.

For m=i=1npiri with n3 and ri1 for each i, let U={iX|ri=1}, V={iX|ri=2} and W={iX|ri3} where X={1,2,...,n} is the indexing set of m. Note that there exists Ti, 1iTAm, twin classes in Am including trivial and non-trivial twin classes. Using these notations, we give the next result on the number of twin classes in Am where m=i=1npiri with ri1, denoted by TAm.

Lemma 2.5

For m=i=1npiri with n3 and ri1 , then the number of twin classes of Am is given as,

TAm=k=0n|U|(n|U|k)2nk1,

where U={iX|ri=1} and X={1,2,...,n} is an index set.

Proof

For m=i=1npiri with n3 and ri1, let X={1,2,...,n} be an indexing set and U consists of first t indices of primes with power one in the representation of m. We first consider the twin classes of the form j=1lpij with 1ln. Notice that there are (nl) classes of the form j=1lpij. As 1ln so there are l=1n(nl) twin classes. Further, we consider the twin classes of the form j=t+1spijαij with at least one 2αijrij and t+1sn. Notice that there are (ntk) classes of this form. As 1knt so in all there are k=1nt(ntk) twin classes. Also, there are (ntk)(2nk1) classes of the form j=1hpijj=t+1spijαij for 1ht, t+1sn, 2αijrij. So, in all there are k=1n|U|(n|U|k)2nk twin classes. Hence, summing up we have the total count of twin classes follows as TAm=k=0n|U|(n|U|k)2nk1. □

For a composite number m=i=1npiri with n3 and ri1 for each i, let W={iX|ri2}, W={iX|ri3}. Note that for W=, the graph Am has no non-trivial twin class. Now for W, let Li={l1,l2,,li}W, Wi=WLi and mi=iWpirikWipk. Using these notations, we give the next result on the number of non-trivial twin classes in Am.

Theorem 2.6

For m=i=1npiri with n3 and ri1 for each i, let W={iX|ri2} , W={iX|ri3} and W . The number of non-trivial twin classes of Am can be found by the formula

i=1|W||V(Ami)|(|W|i)+2|W|1,

where mi=iWpirikWipk , Wi=WLi and Li={l1,l2,,li}W .

Proof

Note that for every nonempty subset Li of W, the set defined as: jLiPj ={jLipjαj|2αjrj} is a non-trivial class of false twins in Am and the possible such classes are 2|W|1. Now for LiW and mi=iWpirikWipk, for each xV(Ami) the set defined as: xjLiPj={xjLipjαj|2αjrj} is a non-trivial class of false twins in Am. Moreover, there are |V(Ami)|(|W|i) such non-trivial false twin classes in Am. Hence, the total number of non-trivial twin classes in Am is i=1|W||V(Ami)|(|W|i)+2|W|1. □

3. Symmetries in the arithmetic graphs

In this section, we describe symmetries of the arithmetic graphs Am and give conditions which induce automorphisms of Am. We begin by proving that for each piP and θAut(Am) then θ(pi)P.

Lemma 3.1

Letm=i=1npiriwithri1andn3. IfθAut(Am), thenθ(P)=PwhereP={p1,p2,...,pn}.

Proof

Note that pix for all xXn. Since θ is automorphism so θ(pi)θ(x) for all xXn. If θ(pi)X1 then θ(pi) can not be adjacent with any vertex of Xn. Hence θ(pi)X1. Further if θ(pi)P then deg(pi)deg(θ(pi)) which is not possible. Hence θ(pi)P. □

In the next result, we discuss symmetries of m and prove that for each xXl where 1ln and θAut(Am), θ(x)Xl.

Lemma 3.2

LetθAut(Am)form=i=1npiri, thenθ(Xl)=Xl, for each1ln.

Proof

Let xXl with l2 and Sx={p1,p2,...,pl}. By definition of the arithmetic graph pix for each i{1,2,...,l}. Since θ is an automorphism so θ(x)θ(pi) for each i{1,2,,l}. Lemma 3.1 gives that θ(pi)P which yields that θ(x) has l distinct primary factors and hence θ(x)Xl. As θ(Xl)=Xl for 2ln, this yields that θ(X1)=X1. □

For a given m=i=1npi, we define the set of primes P={p1,p2,...,pn} in the canonical representation of m. We denote by Sym(P), the symmetric group of P, consisting of all the permutations on the set P. Note that Sym(P)Sn where Sn is symmetry group on n elements. For a permutation η on P, we associate an automorphism θ on x=p1p2...pjV(Am) where jn by

θ(x)=η(p1)η(p2)....η(pj), (1)

and |Sx|=j;1jl for 1ln.

Theorem 3.3

For m=i=1npi , SnAut(Am) .

Proof

Let ηSym(P) and θAut(Am). We define a function φ:Sym(P)Aut(Am) as φ(η)=θ and prove that φ is a group isomorphism.

(i) φ is well defined: Let η1,η2Sym(P) be two permutations such that η1η2. Let φ(η1)=θ1 and φ(η2)=θ2. Since η1η2, therefore η1(pi)η2(pi), for at least one i, 1il, yields {η1(p1),η1(p2),η1(p3),...,η1(pl)} {η2(p1),η2(p2),η2(p3),...,η2(pl)} Sθ1(x)Sθ2(x)θ1(x)θ2(x) θ1θ2.

(ii)φ is injective: Let θ1,θ2Aut(Am) such that θ1θ2. Let η1,η2Sym(P) be two permutations such that φ(η1)=θ1 and φ(η2)=θ2. Since, θ1θ2, therefore there exists a vertex xXl, for some 1ln such that θ1(x)θ2(x). Let Sx={p1,p2,p3,...,pl}, then Sθ1(x)Sθ2(x){η1(p1),η1(p2),η1(p3),...,η1(pl)} {η2(p1),η2(p2),η2(p3),...,η2(pl)}η1(pi)η2(pi), for at least one i, 1il,η1η2.

(iii)φ is surjective: For any automorphism θAut(Am), there exists a permutation ηSym(P) such that φ(η)=θ.

(iv)φ is homomorphism: Let η1,η2Sym(P) and θ1,θ2Aut(Am) such that φ(η1)=θ1 and φ(η2)=θ2. We claim that φ(η1η2)=φ(η1)φ(η2)=θ1θ2. Let φ(η1η2)=θ and xXl, for some l, 1ln such that Sx={p1,p2,p3,...,pl}. Then θ(x) has skeleton Sθ(x)={η1η2(p1),η1η2(p2),η1η2(p3),...,η1η2(pl)} ={η1(η2(p1)),η1(η2(p2)),η1(η2(p3)),...,η1(η2(pl))}=Sθ1(θ2(x)) =Sθ1θ2(x)θ=θ1θ2φ(η1η2)=φ(η1)φ(η2).

Thus, φ is an isomorphism between Sym(P) and Aut(Am). □

In the next result, we give the automorphisms of Am for m=i=1npiri with ri1 and n2 by using permutations of {p1,p2,,pn}.

Theorem 3.4

Let m=i=1npiri with n2 and ηSym(P) then θ:V(Am)V(Am) defined as θ(p1α1p2α2...pnαn)=η(p1)α1η(p2)α2...η(pn)αn for each p1α1p2α2...pnαnV(Am) , where 0αiri , i with at least one αi0 for some i, is an automorphism of Am .

Proof

Let θ:V(Am)V(Am) be a function defined as θ(p1α1p2α2...pnαn)=η(p1)α1η(p2)α2...η(pn)αn for each p1α1p2α2...pnαnV(Am) and we prove that θ is an automorphism of Am. Let x=p1α1p2α2...pnαn and y=p1β1p2β2...pnβn with 0αi,βiri for each i and at least one αi0 and one βi0 for some 1in.

(i) θ is well defined: For x=y, we have αi=βi and ηSym(P) gives that θ(x)=θ(y).

(ii) θ is injective: Suppose θ(x)=θ(y) η(p1)α1η(p2)α2...η(pn)αn=η(p1)β1η(p2)β2...η(pn)βn gives that αi=βi for each i so x=y.

(iii) θ is surjective: For v=p1α1p2α2...pnαnV(Am) with piP and ηSym(P). Let piP such that η(pi)=pi for each 1in then u=p1α1p2α2...pnαnV(Am) such that θ(u)=v.

(iv) θ is homomorphism: Next, we assume xy then x and y have different parity and gcd(x,y)=pi for some 1in and SxSy=pi.

Clearly, η(pi)=Sθ(x)Sθ(y). Now gcd(x,y)=pi gives that αi=1 or βi=1 or αi=βi=1 which gives that gcd(θ(x),θ(y))=η(pi). Hence, θ(x)θ(y).

Now, suppose that xy, then we consider following cases:

Case 1. If x and y have the same parity then Sx=Sy and ηSym(P) gives that Sθ(x)=Sθ(y) which implies θ(x)θ(y).

Case 2. If SxSy= then Sθ(x)Sθ(y)= because ηSym(P) which yields that θ(x)θ(y).

Hence, θ is an automorphism of Am. □

In the following corollary, using equation (1) and above theorem, we give the form of θ(x)=iSxη(pi) for all xV(Am) if θ is an automorphism.

Corollary 3.5

Form=i=1npiwithn3, a mappingθ:V(Am)V(Am)is an automorphism ofAmif and only if there exists a permutationηSym(P)such that θ has the formθ(x)=iSxη(pi)for allxV(Am).

4. Fixing number of the arithmetic graphs

In this section, we will give results on the fixing number of the arithmetic graph Am for a given m. For a given m=i=1npiri with all ri1 and n2 with indexing set X={1,2,3,,n}, let U={iX|ri=1}, V={iX|ri=2} and W={iX|ri3} be the collection of subsets of X which give a partition of X. Note that for n=2, at least one of U, V and W is empty set. Notice that for m=i=12piri with ri1, p1,p2 are adjacent with all vertices of the form i=12piαi with 1αiri. This yields that N(p1)=N(p2) and deg(p1)=deg(p2)=r1r2 which implies that Am has non-trivial symmetries and fix(Am)1. This motivates us to study the fixing number of the arithmetic graph for m=i=12piri where ri1 in the following theorem.

Theorem 4.1

For a composite number m=i=12piri with ri1 , we have

fix(Am)={1ifU,W=2ifU=,V,W=1+2(ri2)ifU,V=,iW1+3(ri2)ifU=,V,iW2i=12(ri2)+i=12(ri1)ifU==V,W.

Proof

We consider the following cases:

(i) For U,V==W, we have m=p1p2 and Am is isomorphic to P3 so fix(Am)=1.

Also, for U, V and W=, the arithmetic graph Am has order 5 and p1,p2 are false twins, moreover, degrees of the remaining three vertices are different which yields that {p1} is a minimum fixing set for Am.

(ii) For U=,V,W=, we have m=p12p22 where the order of Am is 8. We claim that F={p1,p12} is a fixing set for Am. We also note that p1,p2 are false twins in Am and N(p12)={p1p2,p1p22} and N(p22)={p1p2,p12p2} which implies that θ(p12)=p22, θ(p22)=p12 and θ(p12p2)=p1p22, θ(p1p22)=p12p2 which shows that p12 can be mapped onto p22 if p1p22 mapped onto p12p2 and vice versa and hence fix(Am)2. For this, we only need to prove that the set F={p1,p12} is a fixing set for Am. Consider θAut(Am) such that θ(s)=s for each sF, we will prove that it is an identity automorphism. Assume contrary that θ is a non-trivial, i.e. there exist xyV(Am) such that θ(x)=y. Note that V(Am)F={p2,p22,p1p2,p12p2,p1p22,p12p22} and deg(p2)=r1r2, deg(p22)=r1, deg(p1p2)=r1+r2, deg(p12p2)=1+r2, deg(p1p22)=r1+1 and deg(p12p22)=2 which yields that θ acts trivially on V(Am). Hence, fix(Am)=2.

(iii) For U,V= and iW without loss of generality suppose m=p1p2r2 with r23. Now the sets {p1,p2}, {p2β|2βr2} and {p1p2β|2βr2} are non-trivial false twin classes in Am and hence fix(Am)1+2(r22). Now we only need to prove that the set F={p1,p23,...,p2r2,p1p23,..,p1p2r2} is a fixing set for Am. Consider θAut(Am) such that θ(s)=s for each sF, we will prove that it is an identity automorphism. Assume contrary that θ is a non-trivial, i.e. there exist xyV(Am) such that θ(x)=y. Note that V(Am)F={p2,p22,p1p2,p1p22} and deg(p2)=r2, deg(p22)=1, deg(p1p2)=1+r2, deg(p1p22)=2 which yields that θ acts trivially on V(Am). Hence, fix(Am)=1+2(r22).

(iv) For U=,V and W without loss of generality suppose m=p12p2r2 with r23. Now the sets {p1,p2}, {p2β|2βr2}, {p1p2β|2βr2} and {p12p2β|2βr2} are non-trivial false twin classes in Am and hence fix(Am)1+3(r22). Now we only need to prove that the set F={p1,p23,,p2r2,p1p23,, p1p2r2,p12p23,, p12p2r2} is a fixing set for Am. Consider θAut(Am) such that θ(s)=s for each sF, we will prove that it is an identity automorphism. Assume contrary that θ is a non-trivial, i.e. there exist xyV(Am) such that θ(x)=y. Note that V(Am)F={p2,p22,p1p2,p12p2,p1p22,p12p22} and deg(p2)=2r2, deg(p22)=2, deg(p1p2)=2+r2, deg(p1p22)=3, deg(p12p2)=1+r2, deg(p12p22)=2 which yields that θ acts trivially on V(Am). Hence, fix(Am)=1+3(r22).

(v) For U=V=, we have m=p1r1p2r2 with r1,r23 and the sets {p1,p2},{p1α|2αr1},{p2β|2βr2}, {p1αp2|2αr1},{p1p2β|2βr2},{p1αp2β|2αr1,2βr2} are non-trivial false twin classes in Am. Therefore, fix(Am)1+2(r12)+2(r22)+(r11)(r21)1 and we need to prove that the set F={p1,p13,...,p1r1,p23,...,p2r2,p1p23,...,p1p2r2,p13p2,...,p1r1p2,p12p23,...,p1r1p2r2} is a fixing set for Am. V(Am)F={p2,p12,p22,p1p2,p1p22,p12p2,p12p22}, where deg(p12)=deg(p22) and deg(p12p2)=deg(p1p22). All other vertices in V(Am)F have distinct degrees. Let θAut(Am) such that θ(s)=s for all sF then we have two cases:

Case 1. Suppose θ(p12)=p22 now θ(p1p23)=p1p23 and p22p1p23 but p12p1p23, hence for being θ an automorphism of Am, we have θ(p12)=p12 and θ(p22)=p22.

Case 2. Suppose θ(p1p22)=p12p2, also θ(p13)=p13 and p13p1p22 but note that θ(p1p22)θ(p13) hence for θ be an automorphism θ(p1p22)=p1p22.

Concluding above discussion, θ is an identity automorphism hence F is a fixing set and fix(Am)=2i=12(ri2)+i=12(ri1). □

From Theorem 4.1, note that for m1=p1p2 and m2=p1r1p2r2 with at least one ri2, 1i2 then fix(Am1)=1 and fix(Am2)2 except for m2=p1p22,p12p2 as its fix(Am2)=1. Therefore, we have the following result.

Corollary 4.2

For m1=p1p2 and m2=p1r1p2r2 with at least one ri2 and 1i2 , then fix(Am1)fix(Am2) .

In the following Fig. 1, the arithmetic graph for m=p1p2p32 is shown.

Figure 1.

Figure 1

An arithmetic graph when m=p1p2p32.

In the following theorem, the fixing number is given for m=i=1npiri, with n3 and W=.

Theorem 4.3

For m=i=1npiri , with n3 and W= , we have

fix(Am)={n1ifU,V=2n|U|3ifU,V2n2ifU=,V.

Proof

(i) For U and V=, we have m=i=1npi. To prove that fix(Am)=n1, we first prove that F={p1,p2,,pn1} is a fixing set for Am. Let θAut(Am) such that θ(pi)=pi for 1in1 as by Lemma 3.1, θ(pn)=pn. Now suppose that θ(x)x for some xV(Am)X1 then SxSθ(x), because there exists at least one piX1 such that piSx and piSθ(x). Now piSx and xV(Am)X1 so xpi but θ(x)θ(pi)=pi because piSθ(x) which gives that θ is not an automorphism. Hence, θ(x)=x for each xV(Am) and F is a fixing set for Am.

Now suppose that FV(Am) such that |F|=n2, then |FX1|n2 then there exists an automorphism θ in Aut(Am) such that θ(pn1)=pn and θ(pn)=pn1 which yields that F is not a fixing set for Am. Hence, fix(Am)=n1.

(ii) For U and V, suppose that |U|=t>0 and V=nt>0. In particular suppose that m=i=1tpij=t+1npj2, where set P1={p1,p2,...,pt}, P2={pt+1,pt+2,...,pn} and P22={pt+12,pt+22,...,pn2} such that X1=P1P2P22. We first prove that F=X1{pt,pn,pn2} is a fixing set for Am. Let θAut(Am) such that θ(pi)=pi for 1it1, it, θ(pj)=pj and θ(pj2)=pj2 for t+1jn1. Note that by Lemma 3.1, θ(P)=P so θ(pt)=pt because deg(pt)deg(pn). By Lemma 3.2, θ(Xl)=Xl for 1ln yields that θ(X1)=X1 so θ(pn)=pn and θ(pn2)=pn2 as deg(pn)deg(pn2). Thus, θ(x)=xxX1.

Now we suppose that θ(x)x and θ(x)=y where x,yXl,2ln. This implies that deg(x)=deg(y) then either SxSy or Sx=Sy. If SxSy then there exist at least one piP such that piSx but piSy and if Sx=Sy then there exists at least one pj2X1 such that pj2x but pj2y so xpi but θ(x)θ(pi)=pi which gives that θ is not an automorphism. Hence, θ(x)=x for each xV(Am) and F is a fixing set for Am.

Now suppose |F|2n|U|4 and |X1F|4 then there exist two vertices x,y in P1,P2 or P22 which yield a non-trivial automorphism and hence F is not a fixing set. Hence, fix(Am)=2n|U|3.

(iii) For U=,V and we have m=i=1npiri. To prove that fix(Am)=2n2. For m=p12p22...pn2, we define the set P2={p1,p2,...,pn} and P22={p12,p22,...,pn2}. Now, we consider the set F={p1,...,pn1,p12,...,pn12}=X1{pn,pn2}V(Am). We first prove that F is a fixing set for Am. Let θAut(Am) such that θ(pi)=pi and θ(pi2)=pi2 for 1in1. Note that as θ(P)=P by Lemma 3.1 and θ(X1)=X1 by using Lemma 3.2 so θ(pn)=pn and θ(pn2)=pn2 as pn and pn2 have different degrees.

Now we suppose that θ(x)x and θ(x)=y where x,yXl,2ln. This implies that deg(x)=deg(y) then either SxSy or Sx=Sy. If SxSy then there exists at least one piP such that piSx but piSy and if Sx=Sy then there exists at least one pi2X1 such that pi2x but pi2y so xpi but θ(x)θ(pi)=pi which gives that θ is not an automorphism. Hence, θ(x)=x for each xV(Am) and F is a fixing set for Am.

Now suppose that |F|2n3 and |X1F|3 then there exist two vertices x,y in P2 or P22 which yield a non-trivial automorphism. Hence, fix(Am)=2n2. □

The following result for the fixing number of the arithmetic graphs when W= is established by using the Theorem 4.3.

Corollary 4.4

Form1=i=1npiandm2=i=1npiriwithn3and1ri2, then the following are true:

(i)fix(Am1)fix(Am2)wherem2has at most oneri=2,

(ii)fix(Am1)<fix(Am2)wherem2has at least threerisequal to 2,

(iii)fix(Am1)=fix(Am2)ifm2has exactly any tworisequal to 2.

Let TAm be number of twin classes and T1,T2,...,TTAm be the twin classes in Am. We consider Ti=Ti{ui} for some fixed uiTi and define F1=i=1ξTi. Note that |F1|=|V(Am)|TAm. For W={iX|ri3}, we define Wtj={iW|ri=tj}, 3tj and 1jk where tk is the highest power of any prime in decomposition of m. Wt1,Wt2,...,Wtk form a partition of W and for 1jk, we define Ftj={pα|αWtjβ} where β is some fixed index in Wtj. Let F2=j=1kFtj and |F2|=j=1k(|Wtj|1). Using the definitions of F1 and F2 and their cardinalities, we give formula for the fixing number of Am where m=i=1npiri, with n3, ri1 and W.

It was established in [28] that for given graph G with TG twin classes, fix(G)|V(G)|TG. In the following theorem, we prove that this bound is attained for m=i=1npiri with 3r1<r2<...<rn. Further, for W, we present exact formula for the fixing number of Am for different cases of U and V.

Theorem 4.5

For m=i=1npiri with n3 , ri1 and W , we have

(i) fix(Am)=j=1k(|Wtj|1)+|V(Am)|TAm if U=,V= ,

(ii) fix(Am)=|U|1+j=1k(|Wtj|1)+|V(Am)|TAm if U,V= ,

(iii) fix(Am)=2|V|2+j=1k(|Wtj|1)+|V(Am)|TAm if U=,V ,

(iv) fix(Am)=2(|U|+|V|)|U|3+j=1k(|Wtj|1)+|V(Am)|TAm if U,V .

Proof

We will prove that every fixing set of Am for W must contains the set F=F1F2. Note that all vertices of non-trivial twin classes belong to F except a fixed xiTi. As all vertices of F are mapped onto themselves so θ(xi)=xixiTi. Hence, F1 fixes every vertex of the non-trivial twin classes for any choice of m. Now, we need to fix vertices of Am to destroy any possible symmetries induced by trivial twin classes which are vertices of the form j=1kpijαij with 1αij2. Now we have following four cases:

(i) Case 1. For U= and V= and |Wtj|=1j,1jk, degrees of all the vertices in trivial twin classes are different which yields that the vertices in the trivial twin classes are fixed which implies that fix(Am)|V(Am)|TAm and by using fix(G)|V(G)|TG, result is established with j=1k(|Wtj|1)=0.

Case 2. Now, for at least one |Wtj|2, then j=1k(|Wtj|1)0. This means that there exist vertices of the form pi and pj with the same degree and by Lemma 3.1, there exist θAut(Am) such that θ(pi)=pj. By fixing F2 with cardinality j=1k(|Wtj|1), all such automorphisms are destroyed. Note that the other vertices of trivial twin classes are of the form pi1pi2...pik with 1kn are fixed as all the vertices of P in Am are fixed by using the same arguments as in Theorem 4.3(i). Hence, fix(Am)=j=1k(|Wtj|1)+|V(Am)|TAm.

(ii) For m=i=1npiri with n3, ri1, U,V= and W, we define P1={p1,p2,...,pt} and |U|=t. Consider the set F={p1,p2,...,pt1}F1F2. We will prove that F is a fixing set for Am. As F1F2F so fix(Am)j=1k(|Wtj|1)+|V(Am)|TAm by using the same arguments as in Theorem 4.5(i). Please note that {p1,p2,...,pt1} fixes all vertices which have p1,...,pt in their canonical representation and {p1,p2,...,pt1}F2 fixes all the vertices of the form pij for 1ijn. Using θ(P)=P and arguments of Theorem 4.3(i), we conclude that vertices of j=1kpijαij with 1kn for some 0αij1 are mapped onto themselves which yields that F is a fixing set for Am. Hence, fix(Am)=|U|1+j=1k(|Wtj|1)+|V(Am)|TAm.

(iii) For m=i=1npiri with n3, ri1, U=,V and W, we define P2={p1,p2,...,ps}, P22={p12,p22,...,ps2} and |V|=s and we consider F={p1,p2,...,ps1}{p12,p22,...,ps12}F1F2. We will prove that F is a fixing set for Am. As F1F2F so fix(Am)j=1k(|Wtj|1)+|V(Am)|TAm by using the same arguments as in Theorem 4.5(i). Please note that {p1,p2,...,ps1}{p12,p22,...,ps12} fixes all the vertices which have p1,...,ps,p12,p22,...,ps2 in their canonical representation. Using θ(P)=P and using arguments of Theorem 4.3(iii), we conclude that vertices of the form j=1kpijαij with 1kn for some 0αij2 are mapped onto themselves which yields that F is a fixing set for Am. Hence, fix(Am)=2|V|2+j=1k(|Wtj|1)+|V(Am)|TAm.

(iv) For m=i=1npiri with n3, ri1, U,V and W, we define P1={p1,p2,...,pt}, P2={pt+1,pt+2,...,ps}, P22={pt+12,pt+22,...,ps2} and |U|=t, |V|=st, we consider a set F consisting of {p1,p2,...,pt1}{pt+1,pt+2,...,ps1}{pt+12,pt+22,...,ps12}F1F2. We will prove that F is a fixing set for Am. As F1F2F so fix(Am)j=1k(|Wtj|1)+|V(Am)|TAm by using the same arguments as in Theorem 4.5(i). Please note that {p1,p2,...,pt1}{pt+1,pt+2,...,ps1}{pt+12,pt+22,...,ps12} fixes all the vertices which have p1,...,pt,pt+1,pt+2,...,ps,pt+12,pt+22,...,ps2 in their canonical representation. Using θ(P)=P and by using arguments of Theorem 4.3(ii), we conclude that vertices of j=1kpijαij for some 0αij2 are mapped onto themselves which yields that F is a fixing set for Am. Hence, fix(Am)=2(|U|+|V|)|U|3+j=1k(|Wtj|1)+|V(Am)|TAm. □

Based on Theorem 4.5, we have the following corollary which gives inequality on the fixing number of the arithmetic graphs associated with m1=i=1npiri and m2=j=1npjsj.

Corollary 4.6

For m1=i=1npiri and m2=j=1npjsj with n3 , then fix(Am1)<fix(Am2) where U,W for m1 and V,W for m2 .

5. Conclusions

Symmetries obtained through automorphisms provide useful insights about a network or a graph. Symmetries can be used to identify relationships among the vertices of graphs. It is important to observe that symmetries of graphs are identified by the similarities of vertices and therefore by identification of symmetries, one can break the symmetries of graph which have vast applications in the field of networking.

In this paper, we have studied symmetries and symmetry-breaking of the arithmetic graphs with some of their associated properties. We have firstly described the properties of the arithmetic graphs such as the distance between vertices, the degree of any vertex and the number of twin classes in arithmetic graphs. We have identified the automorphism group of Am for m=i=1npi and have proved that it is isomorphic to the symmetric group of n elements. Moreover, we have studied the conditions on permutations of {p1,p2,,pn} of the arithmetic graphs which induce automorphisms of Am for m=i=1npiri with ri1. Therefore, it is interesting to examine the non-trivial symmetries of the arithmetic graphs, which lead to investigate the fixing numbers of the arithmetic graphs Am of m=i=1npiri with ri1.

Moreover, we have investigated the twin-free arithmetic graphs with their fixing numbers when m is the form of i=1npi and i=1npiri with 1ri2. We have also noted the difference between the metric dimension and the fixing number of the arithmetic graphs and noticed that for p1p2, p1p2r2,r23, p12p2r2,r23, p1r1p2r2,r1,r23, the metric dimension and the fixing number of Am are equal.

Table 1 shows a comparison of the metric dimension proved by Rehman et al. [42] and the fixing numbers of the arithmetic graphs Am. Please note that the results coincide for some composite numbers m.

Table 1.

Comparison of the fixing number and the metric dimension of m of the arithmetic graphs.

m dim(Am) fix(Am)
i=12pi 1 1
i=12piri,r23 1 + 2(r2 − 2) 1 + 2(r2 − 2)
i=12pi2 3 2
i=12piri,r23 2 + 3(r2 − 2) 2 + 3(r2 − 2)
i=12piri,ri3 2(r1 − 2)+2(r2 − 2)+(r1 − 1)(r2 − 1) 2(r1 − 2)+2(r2 − 2)+(r1 − 1)(r2 − 1)
i=1npi n n − 1
i=1npi2 2n 2n − 2
i=1tpii=t+1npi2 and n ≥ 4 2n − t 2n − t − 3

In [28], “Garijo et al. posed the question that whether the difference between the metric dimension and the fixing number of a graph can be arbitrarily large of order n?” and studied the stated problem. In [42], Rehman et al. computed the metric dimension of the arithmetic graphs and it was interesting to note the difference between the metric dimension and the fixing number of the arithmetic graphs Am. It is an open and interesting question that for m=i=1npiri with n3, ri1 and W, can the difference between the metric dimension and the fixing number of Am be arbitrarily large? The answer seems positive if the trivial twin classes of vertices are increased.

CRediT authorship contribution statement

Aqsa Shah: Conceived and designed the experiments; Performed the experiments; Wrote the paper.

Imran Javaid: Conceived and designed the experiments; Analyzed and interpreted the data.

Shahid Ur Rehman: Contributed reagents, materials, analysis tools or data.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

The authors are grateful to the editor and the anonymous referees for their suggestions and constructive comments which lead to improvement in this paper.

Data availability

Data included in article/supp.material/referenced in article.

References

  • 1.Boutin D.L. The determining number of a Cartesian product. J. Graph Theory. 2009;61:77–87. [Google Scholar]
  • 2.Caceres J., Garijo D., Gonzalez A., Marquez A., Puertas M.L. The determining number of Kneser graphs. Discret. Math. Theor. Comput. Sci. 2013;15:1–14. [Google Scholar]
  • 3.Gibbons C., Laison J. Fixing numbers of graphs and groups. Electron. J. Comb. 2009;26:1–13. [Google Scholar]
  • 4.Bailey R.F., Cameron P.J. Base size, metric dimension and other invariants of groups and graphs. Bull. Lond. Math. Soc. 2011;43:209–242. [Google Scholar]
  • 5.Chvatal V. Mastermind. Combinatorica. 1983;3:325–329. [Google Scholar]
  • 6.Erwin D., Harary F. Destroying automorphisms by fixing nodes. Discrete Math. 2006;306:3244–3252. [Google Scholar]
  • 7.Sebo A., Tannier E. On metric generators of graphs. Math. Oper. Res. 2004;29:383–393. [Google Scholar]
  • 8.Shapiro H., Sodeeberg S. A combinatory detection problem. Am. Math. Mon. 1970;70:1066–1070. [Google Scholar]
  • 9.Brooks J., Carbonero A., Vargas J., Flórez R., Rooney B., Narayan D. Removing symmetry in circulant graphs and point-block incidence graphs. Mathematics. 2021;9 doi: 10.3390/math9020166. [DOI] [Google Scholar]
  • 10.Caceres J., Hernando C., Mora M., Pelayo I.M., Puertas M.L., Seara C., Wood D.R. On the metric dimension of Cartesian products of graphs. SIAM J. Discrete Math. 2007;21:423–441. [Google Scholar]
  • 11.Singh P., Sharma S., Sharma S.K., Bhat V.K. Metric dimension and edge metric dimension of windmill graphs. AIMS Math. 2021;6:9138–9153. [Google Scholar]
  • 12.Ali A., Raza Z., Bhatti A.A. On the augmented Zagreb index. Kuwait J. Sci. 2016;43:123–138. [Google Scholar]
  • 13.Ali A., Bhatti A.A., Raza Z. The augmented Zagreb index, vertex connectivity and matching number of graphs. Bull. Iranian Math. Soc. 2016;42:417–425. [Google Scholar]
  • 14.Raza Z., Bhatti A.A., Ali A. More on comparison between first geometric-arithmetic index and atom-bond connectivity index. Miskolc Math. Notes. 2016;17:561–570. [Google Scholar]
  • 15.Boutin D.L. Identifying graph automorphisms using determining sets. Electron. J. Comb. 2006;13 doi: 10.37236/1104. [DOI] [Google Scholar]
  • 16.Benish H., Javaid I., Murtaza M. Automorphism related parameters of a graph associated to a finite vector space. Util. Math. 2019;111:27–33. [Google Scholar]
  • 17.P.J. Cameron, Automorphisms of graphs, Queen Mary, University of London, E1 4NS, U.K., 2001.
  • 18.Kutnar K., Petecki P. On automorphisms and structural properties of double generalized Petersen graphs. Discrete Math. 2016;339:2861–2870. [Google Scholar]
  • 19.Ou S., Wang D., Tian F. The automorphism group of zero-divisor graph of a finite semisimple ring. Commun. Algebra. 2020;48:2388–24050. [Google Scholar]
  • 20.Wang L. Automorphisms of the zero-divisor graph of the ring of all n×n matrices over a finite field. Discrete Math. 2016;339:2036–2041. [Google Scholar]
  • 21.Zhang H., Nan J., Tang G. Automorphisms of the zero-divisor graph of 2×2 matrix ring over Zps. J. Algebra Appl. 2017;16 [Google Scholar]
  • 22.Zhou J.X., Ghasemi M. Automorphisms of a family of cubic graphs. Algebra Colloq. 2013;20:495–506. [Google Scholar]
  • 23.Albertson M.O., Collins K.L. Symmetry breaking in graphs. Electron. J. Comb. 1996;3 doi: 10.37236/1242. [DOI] [Google Scholar]
  • 24.Albertson M.O., Boutin D.L. Using determining sets to distinguish Kneser graphs. Electron. J. Comb. 2007;14 doi: 10.37236/938. [DOI] [Google Scholar]
  • 25.Harary F., Melter R.A. On the metric dimension of a graph. Ars Comb. 1976;2:191–195. [Google Scholar]
  • 26.Slater P.J. Proceedings of the 6th Southeastern Conference on Combinatorics, Graph Theory, and Computing. vol. 14. 1975. Leaves of trees; pp. 549–559. (Congr. Numer.). [Google Scholar]
  • 27.Caceres J., Garijo D., Puertas M., Seara C. On the determining number and the metric dimension. Electron. J. Comb. 2010;17 doi: 10.37236/335. [DOI] [Google Scholar]
  • 28.Garijo D., Gonzalez A., Marquez A. The difference between the metric dimension and the determining number of a graph. Appl. Math. Comput. 2014;249:487–501. [Google Scholar]
  • 29.González A., Puertas M.L. Removing twins in graphs to break symmetries. Mathematics. 2019;7 [Google Scholar]
  • 30.Das A. 12 CALDAM 2020. vol. 12016. Springer; 2020. Determining number of generalized and double generalized Petersen graph; pp. 131–140. (LNCS). [Google Scholar]
  • 31.Pan J., Guo X. The full automorphism groups, determining sets and resolving sets of coprime graphs. Graphs Comb. 2019;35:485–501. [Google Scholar]
  • 32.Ou S., Wong D., Tian F., Zhou Q. Fixing number and metric dimension of a zero-divisor graph associated with a ring. Linear Multilinear Algebra. 2021;69:1789–1802. [Google Scholar]
  • 33.Alon N., Erdos P. An application of graph theory to additive number theory. Eur. J. Comb. 1985;6:201–203. [Google Scholar]
  • 34.Chaluvaraju B., Chaitra V. Sign domination in arithmetic graphs. Gulf J. Math. 2016;4:49–54. [Google Scholar]
  • 35.Somer L., Krizek M. On a connection of number theory with graph theory. Czechoslov. Math. J. 2004;54:465–485. [Google Scholar]
  • 36.Vasumathi N. S. V. University; Tirupati, India: 1994. Number theoretic graphs. Ph. D. Thesis. [Google Scholar]
  • 37.Jenitha L.M., Sujitha S. Super connected and hyper connected arithmetic graphs. Malaya J. Mat. 2020;1:243–247. [Google Scholar]
  • 38.Maheswari S.U., Maheswari B. Some domination parameters of the arithmetic graph Vn. IOSR J. Math. 2012;2:14–18. [Google Scholar]
  • 39.Sujitha S., Jenitha L.M. Super edge connectivity number of an arithmetic graph. J. Indones. Math. Soc. 2022;28:69–74. [Google Scholar]
  • 40.Rao S., Sreenivansan K.V. The split domination in arithmetic graphs. Int. J. Comput. Appl. 2011;29:46–49. [Google Scholar]
  • 41.Vasumathi N., Vangipuram S. The annihilator domination in some standard graphs and arithmetic graphs. Int. J. Pure Appl. Math. 2016;106:123–135. [Google Scholar]
  • 42.Rehman S.U., Imran M., Javaid I. On the metric dimension of the arithmetic graph of a composite number. Symmetry. 2020;12:607. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data included in article/supp.material/referenced in article.


Articles from Heliyon are provided here courtesy of Elsevier

RESOURCES