Skip to article sections

Let H be a connected subgraph of a connected graph G. The H-structure connectivity of the graph G⁠, denoted by κ(G;H), is the minimum cardinality of a minimal set of subgraphs F={H1′,H2′,…,Hm′} in G⁠, such that every H′i∈F is isomorphic to H and removal of F from G will disconnect G⁠. The H-substructure connectivity of the graph G⁠, denoted by κs(G;H), is the minimum cardinality of a minimal set of subgraphs F={J1′,J2′,…,Jm′} in G⁠, such that every Ji′∈F is a connected subgraph of H and removal of F from G will disconnect G. In this paper, we provide the H-structure and the H-substructure connectivity of the circulant graph Cir(n,Ω) where Ω={1,…,k,n−k,…,n−1},1≤k≤⌊n2⌋ and the hypercube Qn for some connected subgraphs H.

A simple graph G=(V,E) is a finite nonempty set V(G) of objects called vertices together with a (possibly empty) set E(G) of unordered pairs of distinct vertices of G called edges. Two distinct vertices u,v∈V(G) are said to be adjacent in G if u and v are connected by an edge and it is represented by {u,v}∈E(G)⁠. A graph G is said to be trivial if it contains only one vertex and no edges. The connectivity is an important indicator of the reliability and fault tolerability of a network. The vertex connectivity of a connected graph G⁠, denoted by κ(G)⁠, is the minimum cardinality of a vertex subset S⊆V(G)⁠, whose removal would disconnect G or G\S is the trivial graph. As a generalization of the vertex connectivity κ(G)⁠, Cheng-Kuan Lin et al. [6] introduced two new kinds of connectivity, called structure connectivity and substructure connectivity. A set F of connected subgraphs of a graph G is a subgraph-cut of G if G\V(F) is a disconnected graph or K1⁠. Let H be a connected subgraph of G⁠. Then F is an H-structure cut if F is a subgraph cut, and every element in F is isomorphic to H⁠. The H -structure connectivity of G⁠, denoted by κ(G;H)⁠, is defined to be the minimum cardinality of all H-structure cuts of G⁠. F is an H-substructure cut if F is a subgraph-cut, and every element in F is isomorphic to a connected subgraph of H⁠. The H-substructure connectivity of G⁠, denoted by κs(G;H)⁠, is the minimum cardinality of all H-substructure cuts of G⁠. Since every H-structure cut is an H-substructure cut κs(G;H)≤κ(G;H)⁠. If H=K1 then we have κ(G;H)=κs(G;H)⁠.

The vertex connectivity κ(G)≥κs(G;H) for every subgraph H of G whereas the relation between vertex connectivity and structure connectivity depends on H⁠. For the graph G⁠, given in Figure 1, κ(G)=2⁠, the structure connectivity of G with respect to the cycle of length 5,κ(G;C5)=1 and the structure connectivity of G with respect to the cycle of length 4,κ(G;C4)=3⁠.

Let Γ be a finite group with e as the identity. A generating set of Γ is a subset Ω such that every element of Γ can be expressed as a product of finitely many elements in Ω⁠. Assume that e∉Ω and a∈Ω implies a−1∈Ω and such a subset Ω is called as a symmetric generating set of Γ⁠. Hereafter, we assume that Ω is a symmetric generating set of a finite group Γ⁠. A Cayley graph is a graph G=(V,E)⁠, where V(G)=Γ and two vertices x and y are adjacent if xy−1∈Ω and it is denoted by Cay(Γ,Ω)⁠. The inclusion of the inverse in Ω for every element of Ω means that Cay(Γ,Ω) is undirected. Since Ω is a generating set for Γ⁠, Cay(Γ,Ω) is connected and Cay(Γ,Ω) is a regular graph of degree |Ω|⁠. Cayley graphs are extensively dealt in the literature and various authors including Dejter [3], Lakshmivarahan [4], Lee [5], Tamizh Chelvam [8], and Wang [10] have worked on Cayley graphs. For example, one can refer the survey by Tamizh Chelvam and Sivagami [9] for domination in Cayley graphs. The Cayley graph constructed out of the finite cyclic group ℤn, n≥2 along with a symmetric generating set Ω is called a circulant graph and the same is denoted by Cir(n,Ω)⁠. The hypercube Qn is the Cayley graph defined on the group ℤ2n with the standard orthonormal basis as the generating set. Cheng-Kuan Lin et al. [6] have obtained κ(Qn;H) and κs(Qn;H) for H∈{K1,K1,1,K1,2,K1,3,C4}⁠. Here, we provide an example in Figure 2, to exhibit a structure cut of the circulant graph Cir(10,{1,2,3,7,8,9}) with respect to K3⁠. In Figure 2, the structure cut is indicated by the dotted lines and note that κs(Cir(10,{1,2,3,7,8,9};K3))=2⁠.

Throughout this paper, G−X denotes the removal of a set X of subgraphs from the graph G and G\B denotes the removal of the set B⊆V(G) from the graph G⁠. When X={H}⁠, G−X is simply denoted as G−H⁠. For a graph G⁠, the open neighborhood N(v) of a vertex v∈V(G) is the set of all vertices which are adjacent to v⁠. The closed neighborhood of v is N[v]=N(v)∪{v}⁠. Let G1=(V1,E1) and G2=(V2,E2) be two graphs. Then the intersection of G1 and G2⁠, denoted by G1∩G2⁠, is the graph whose vertex set is V1∩V2 and the edge set is E1∩E2⁠. The union of two disjoint vertex sets graphs G1=(V1,E1) and G2=(V2,E2)⁠, denoted by G1∪G2⁠, is the graph whose vertex set is V1∪V2 and the edge set is E1∪E2⁠. For basic definitions and properties related to graph theory, one can refer [2]. For undefined definitions related to algebraic graph theory, one can refer [1].

In Section 2, we obtain the H-structure and the H-substructure connectivity of the circulant graph Cir(n,Ω) where Ω={1,…,k,n−k,…,n−1}⁠, for 1≤k≤⌊n2⌋ with respect to some of its connected subgraphs. For integers n(≥5) and m with 2≤m≤n−2⁠, Mane [7] proved that κ(Qn;Ck)≤n−m⁠, where k is a positive even integer with 2m<k<2m+1 and observed that κ(Q4;C6)=2⁠. In Section 3, for n≥4⁠, we obtain the exact value for κ(Qn;C6)⁠.

Throughout this section n(≥2)⁠, m and k are integers such that 1≤k≤⌊n2⌋ and Ω={1,2,…,k,n−k,…,n−1}⁠. We take the elements of ℤn as ℤn={0,1,…,n−1}⁠. The following result due to Wang [10] is useful in the paper.

Theorem 2.1

([10, Wang]). Let n and k be positive integers such that 1≤k≤⌊n2⌋, Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). Then κ(G)=|Ω|.

By definition and from Theorem 2.1, we have the following corollary.

Corollary 2.2.

Let n(≥2) and k be positive integers such that 1≤k≤⌊n2⌋, Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). Then κ(G;K1)=|Ω| and κs(G;K1)=|Ω|.

A star graph K1,m(m≥1) is a complete bipartite graph comprised of two partite sets of vertices of sizes 1 and m respectively, such that two vertices are adjacent if and only if they are in different partite sets. If K1,m is a subgraph of Cir(n,Ω)⁠, then m≤|Ω|⁠. If k=⌊n2⌋ and Ω={1,…,k,n−k,…,n−1}⁠, then the circulant graph G=Cir(n,Ω) is the complete graph Kn⁠. Now, we obtain the structure connectivity of Kn⁠, as a circulant graph with respect to K1,m⁠. If m+1 does not divide n−1⁠, then removal of λK1,m does not disconnect Cir(n,Ω) for any λ⁠. Hence the structure connectivity of Kn with respect to K1,m is meaningful only when m+1 divides n−1⁠.

Lemma 2.3.

Let n(≥2) and k be positive integers such that k=⌊n2⌋, Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). For a positive integer m with m≤n−2, κs(G;K1,m)=⌈n−1m+1⌉. Also, κ(G;K1,m)=n−1m+1 if m+1 divides n−1.

Proof.

By the assumption on n,k and Ω⁠, Cir(n,Ω)=Kn⁠. By Theorem 2.1, G is (n−1)- connected. Let F be a K1,m-substructure cut with minimum cardinality of G=Cir(n,Ω)⁠. Suppose κs(G;K1,m)<⌈n−1m+1⌉⁠. Then |V(F)|<n−1 and G\V(F) is disconnected, which is a contradiction to G is (n−1)-connected. Hence κs(G;K1,m)≥⌈n−1m+1⌉⁠.

For 1≤i≤⌈n−1m+1⌉−1⁠, let Hi be the subgraph of G with i(m+1) as the central vertex and i(m+1)−1,…,i(m+1)−m as the end vertices and hence isomorphic to K1,m⁠. Consider the subgraph H⌈n−1m+1⌉ of G with (n−1) as the central vertex and all remaining vertices of (G−{H1,…,H⌈n−1m+1⌉−1})\{0,n−1} as end vertices. Note that H⌈n−1m+1⌉ is isomorphic to a subgraph of K1,m⁠. Clearly G−{H1,…,H⌈n−1m+1⌉} is the trivial graph K1⁠. Hence κs(G;K1,m)≤⌈n−1m+1⌉⁠. Thus κs(G;K1,m)=⌈n−1m+1⌉⁠.

Suppose m+1 divides n−1⁠. As mentioned above in the proof, Hn−1m+1 is a subgraph of G isomorphic to K1,m and so κ(G;K1,m)≤n−1m+1⁠. Now n−1m+1=κs(G;K1,m)≤κ(G;K1,m)≤n−1m+1⁠. Hence κ(G;K1,m)=n−1m+1⁠. □

In Lemma 2.3, we have considered k=⌊n2⌋ in which case G=Cir(n,Ω) is complete. Now, we consider k<⌊n2⌋⁠, so that G=Cir(n,Ω) can never be complete. By considering k<⌊n2⌋⁠, we determine the structure and substructure connectivity of Cir(n,Ω) with respect to K1,m where m≤2k⁠.

Theorem 2.4.

Let n(≥4),k and m be positive integers such that 1≤k<⌊n2⌋ and m≤2k. Let Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). Then the following are equivalent:

  • (i) κs(G;K1,m)=1;

  • (ii) m+1=2k+1=n−1;

  • (iii) κ(G;K1,m)=1.

Proof.

Since k<⌊n2⌋⁠, |Ω|=2k<n−1⁠.

(i) ⇒ (ii). Assume that κs(G;K1,m)=1⁠. So that there exists a subgraph K1,t of K1,m for some t,t≤m≤2k such that G−K1,t is disconnected or a trivial graph. Since G is vertex transitive, one can have the central vertex of K1,t as u=0⁠. Consider the subgraph H of G induced by 〈{0,±1,…,±k}〉. The graph G−H is connected and the vertices of H other than 0 are adjacent to either k+1 or n−(k+1) in G⁠. Note that G−H is a subgraph of G−K1,t⁠. Suppose t<2k⁠, then G−K1,t is connected, which is a contradiction. Hence t=m=2k and G−K1,t=G−K1,2k⁠. It is easy to observe that the graphs G−K1,2k and G−H are equal. It is known that m=2k≤n−2⁠. Suppose 2k<n−2⁠, then G−K1,t=G−H is connected, which is a contradiction. This implies that t=m=2k=n−2⁠. Hence m+1=2k+1=n−1⁠.

(ii) ⇒ (iii). Assume that m+1=2k+1=n−1⁠. For u∈V(G)⁠, deg(u)=2k=n−2 and hence G\N[u]=K1⁠. Since |N(u)|=2k=m⁠, K1,m is a subgraph of 〈N[u]〉 and hence removal of 〈N[u]〉 from G is same as removing K1,m from G⁠. Thus κ(G;K1,m)=1⁠.

(iii) ⇒ (i). Since κs(G;K1,m)≤κ(G;K1,m)=1⁠, κs(G;K1,m)=1⁠. □

Remark 2.5.

Let n(≥6) and k be positive integers such that 2≤k<⌊n2⌋⁠, Ω={1,…,k,n−k,…,n−1}⁠, G=Cir(n,Ω) and m≤2k⁠. Even without the condition n>(m+1)⌈2km+1⌉⁠, one can talk about κs(G;K1,m)⁠, whereas it is not so in the case of κ(G;K1,m)⁠. For, if n≤(m+1)⌈2km+1⌉⁠, then for any integer λ with |V(λK1,m)|≤n⁠, removal of λK1,m does not disconnect G⁠.

Consider n≤(m+1)⌈2km+1⌉⁠. If λ<⌈2km+1⌉⁠, then |V(λK1,m)|<2k and hence by Theorem 2.1, G is connected after removal of λK1,m from G⁠. On the other hand if λ≥⌈2km+1⌉⁠, then |V(λK1,m)|≥(m+1)⌈2km+1⌉≥n⁠. This along with |V(λK1,m)|≤n yields |V(λK1,m)|=n⁠. Thus G=λK1,m⁠.

Now, we attempt to obtain κ(Cir(n,Ω);K1,m) and κs(Cir(n,Ω);K1,m)⁠, for 2≤m+1≤k and Ω={1,…,k,n−k,…,n−1}⁠.

Theorem 2.6.

Let n(≥6) and k be positive integers such that 2≤k<⌊n2⌋, Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). If m is an integer such that 2≤m+1≤k and (m+1)⌈2km+1⌉<n, then κ(G;K1,m)=⌈2km+1⌉ and κs(G;K1,m)=⌈2km+1⌉.

Proof.

Let ai=n−(k−i+1) for 1≤i≤k⁠, ai=i−k for k+1≤i≤2k and bj=j for k+1≤j≤n−(k+1)⁠. By division algorithm 2k=(m+1)s+r and k=(m+1)h+r′ for some r and r′ with 0≤r≤m and 0≤r′≤m⁠.

For 1≤i≤s=2k−rm+1⁠, let Hi be defined as follows.

V(Hi)={a(m+1)i−m,a(m+1)i−(m−1),…,a(m+1)i} and edge set

Further when r≠0⁠, let Hs+1 be defined as follows.

V(Hs+1)={v1,…,vm+1:vi={a2k−(r−i)if 1≤i≤rbk+i−rif r+1≤i≤m+1} and edge set

In G⁠, two vertices u and v are adjacent if and only if u,v∈ℤn has the property that |u−v|≤k⁠. Since |a(m+1)i−(m−r′)−a(m+1)i−(m−j)|≤k for every 0≤j≤m and |vr′+1−vj|≤k for every 1≤j≤m+1⁠, Hi is indeed a subgraph of G for every 1≤i≤s+1⁠.

Note that, each Hi is isomorphic to K1,m⁠. Let H be the union of subgraphs given by H={⋃i=1sHi=⋃i=1sK1,mif r=0;⋃i=1s+1Hi=⋃i=1s+1K1,mif r≠0.

Note that V(H)=⌈2km+1⌉K1,m⁠, N(0)⊆V(H)⁠, G−H is disconnected with {0} as one component. Thus, κ(G;K1,m)≤⌈2km+1⌉ and so κs(G;K1,m)≤⌈2km+1⌉⁠. By Theorem 2.1, G is 2k-connected. Suppose there exists a set F′={H1′,…,Ht′} of subgraphs of G such that every Hi′∈F′ is isomorphic to a subgraph of K1,m⁠, t<⌈2km+1⌉ and G−F′ is disconnected. Let X=V(F′)⁠. Clearly |X|<2k and by the assumption G\X is disconnected, which is a contradiction to G is 2k-connected.

Thus κs(G;K1,m)≥⌈2km+1⌉ and so κ(G;K1,m)≥⌈2km+1⌉⁠. Hence κ(G;K1,m)=κs(G;K1,m)=⌈2km+1⌉⁠. □

Now we obtain κ(Cir(n,Ω);K1,m) and κs(Cir(n,Ω);K1,m)⁠, for k<m+1≤2k+1 and Ω={1,…,k,n−k,…,n−1}⁠.

Lemma 2.7.

Let n(≥6) and k be positive integers such that 2≤k<⌊n2⌋, Ω={1,…,k,n−k,…,n−1} and G=Cir(n,Ω). If m is an integer with k<m+1≤2k+1 and n>(m+1)⌈2km+1⌉, then

Proof.

By Theorem 2.6, κ(G;K1,m)=κs(G;K1,m)=2 for m+1=k⁠. This gives that κ(G;K1,m)≤2 and so κs(G;K1,m)≤2 when k<m+1≤2k+1⁠. By Theorem 2.4, κ(G;K1,m)=1=κs(G;K1,m) if and only if m+1=2k+1=n−1⁠. Hence for the other cases κ(G;K1,m)≥2 and κs(G;K1,m)≥2⁠. Thus,

Now we provide an example for the K1,4-substructure connectivity of the circulant graph Cir(16,{1,2,14,15}) in Figure 3. Here n=16,k=2,m=4 and k<m+1⁠. The substructure cut is F={H1≅K1,3,H2≅K1,2}⁠. In Figure 3, the substructure cut F is indicated by the dotted lines.

The n-dimensional hypercube Qn is the Cayley graph defined on the group ℤ2n with generating set as the standard orthonormal basis. Note that Qn contains 2n vertices and n2n−1 edges. Actually two distinct vertices x=(x1x2…xn) and y=(y1y2…yn) in V(Qn) are adjacent if and only if xi≠yi for exactly one i (⁠1≤i≤n⁠). For any vertex x=(x1x2…xn) in Qn⁠, let (x)i=(x1ix2i…xni) where xji=xj for every j≠i and xii=1−xi⁠. Note that {(x)i}1≤i≤n is the neighborhood set of x in Qn⁠. For each t=0,1⁠, we have two (n−1)-dimensional subgraphs Qnt of Qn where V(Qnt)={x|x=(x1x2…xn)∈V(Qn) and xn=t} and E(Qnt)={{x,y}|{x,y}∈E(Qn) and x,y∈V(Qnt)}⁠. Obviously, Qnt is isomorphic to Qn−1 for each t=0,1⁠. The path Pm of length m is a walk with m+1 distinct vertices and m distinct edges. The cycle Cm of length m is a closed path that contains m distinct vertices.

Cheng-Kuan Lin et al. [6] proved the following theorem for the substructure connectivity of hypercube Qn with respect to the cycle C4⁠.

Theorem 3.1

([6, Theorem 10]). For n≥4⁠, κs(Qn;C4)=⌈n2⌉⁠.

For integers n(≥5),k and m with k is a positive even integer, 2m<k<2m+1 and 2≤m≤n−2⁠, Mane [7] considered the substructure connectivity of Qn with respect to the cycle C6⁠. In fact Mane [7] proved that κ(Qn;Ck)≤n−m⁠, and κ(Q4;C6)=2⁠. In this section, for n≥4⁠, we obtain the exact value for κ(Qn;C6)⁠.

First, we obtain the structure connectivity and the substructure connectivity of hypercube Qn with respect to P3⁠, the path of length 3.

Corollary 3.2.

For n≥4, κ(Qn;P3)=κs(Qn;P3)=⌈n2⌉.

Proof.

By Theorem 3.1, κs(Qn;C4)=⌈n2⌉⁠. Since all subgraphs of P3 are also subgraphs of C4⁠, we have κs(Qn;P3)≥⌈n2⌉⁠.

For 1≤i≤⌈n2⌉−1⁠, consider the paths of length 3⁠, Ri:(ai1⋯ain)−(bi1⋯bin)−(ci1⋯cin)−(di1⋯din) where

For odd n⁠, let R⌈n2⌉:(0…01)−(10…01)−(10…011)−(10…010) and for even n⁠, let R⌈n2⌉:(0…010)−(0…011)−(0…01)−(10…01)⁠. The removal of these paths Ri⁠, for 1≤i≤⌈n2⌉⁠, of length 3 disconnects Qn with (0…00) as an isolated vertex. Hence κ(Qn;P3)≤⌈n2⌉⁠.

Thus, we have ⌈n2⌉≤κs(Qn;P3)≤κ(Qn;P3)≤⌈n2⌉ and so κ(Qn;P3)=κs(Qn;P3)=⌈n2⌉⁠. □

In the following lemma, we obtain an upper bound for the structure connectivity of Qn with respect to C6⁠.

Lemma 3.3.

For n≥3, κ(Qn;C6)≤⌈n3⌉.

Proof.

By division algorithm, n=3q+r⁠, 0≤r≤2⁠. For 1≤i≤q⁠, consider the cycles Bi of length 6 given below:

Bi:(ai1⋯ain)−(bi1⋯bin)−(ci1⋯cin)−(di1⋯din)−(ei1⋯ein)−(fi1⋯fin)−(ai1⋯ain) where

If r=1⁠, let Bq+1:(0…001)−(0…011)−(0…0111)−(0…01111)−(00…01101)−(0…01001)−(00…001)⁠.

If r=2⁠, let Bq+1:(0…010)−(0…0110)−(0…0111)−(0…0101)−(0…001)−(0…011)−(00…010)⁠.

The removal of cycles B1,B2,…,B⌈n3⌉ disconnects Qn with (0…00) as an isolated vertex. Hence κ(Qn;C6)≤⌈n3⌉⁠. □

For each n≥3⁠, Z6n is a collection of 6-cles of Qn and the same is taken as {{u,v,w,x,y,z}|{u,v},{v,w},{w,x},{x,y},{y,z},{z,u}∈E(Qn)}⁠. Let Z6n={X1,…,Xm} be a subset of collection of 6-cycles of Qn⁠. For i=0,1⁠, (Z6n)i⊆Z6n is the subgraph ⋃j=1mXj∩Qni of Qn⁠. Cheng-Kuan Lin et al. [6] obtained the substructure connectivity of hypercube Qn with respect to K1,2 and the same the stated below to obtain a lower bound for the substructure connectivity of the hypercube with respect to C6⁠.

Lemma 3.4

([6, Theorem 6]). For n≥3⁠, κ(Qn;K1,2)=⌈n2⌉⁠.

Lemma 3.5.

If |Z64|<2⁠, then Q4−Z64 is connected.

Proof.

If |Z64|=0⁠, then Q4−Z64=Q4⁠, hence is connected. Assume that |Z64|=1 and Z64={C:u1−u2−u3−u4−u5−u6−u1}⁠.

Suppose ui∈Q40 for all 1≤i≤6⁠. Since Q41 is connected and every vertex of Q40−Z64 is connected to a vertex in Q41⁠, we get that Q4−Z64 is connected.

If ui∈Q41 for all 1≤i≤6⁠, then by similar arguments as above, Q4−Z64 is connected.

Assume that V(C)∩V(Q40)≠φ and V(C)∩V(Q41)≠φ⁠. Without loss of generality one can assume that |V(C)∩V(Q40)|≤|V(C)∩V(Q41)|⁠. Then we have two cases.

Case 1.

Let |V(C)∩V(Q40)|=2 and |V(C)∩V(Q41)|=4⁠. Clearly Q40−(Z64)0 is connected and every vertex in Q41−(Z64)1 is adjacent to a vertex in Q40−(Z64)0⁠. Hence Q4−Z64 is connected.

Case 2.

Let |V(C)∩V(Q40)|=3 and |V(C)∩V(Q41)|=3⁠. Note that subgraphs induced by |V(C)∩V(Q40) and |V(C)∩V(Q41) are subgraph isomorphic to K1,2 in Q40 and Q41 respectively. Further both Q40 and Q41 are isomorphic to Q3⁠. By Lemma 3.4, removal of a K1,2 does not disconnect Q40 and Q41⁠. Thus both Q40−(Z64)0 and Q40−(Z64)0 are connected. Also there exists a vertex x∈Q40−(Z64)0 adjacent to (x)4∈Q41−(Z64)1⁠. Thus Q4−Z64 is connected. □

In the following lemma, we obtain a lower bound for the structure connectivity of Qn with respect to C6⁠.

Lemma 3.6.

For an integer n≥4, κ(Qn;C6)≥⌈n3⌉.

Proof.

By induction on n⁠. By Lemma 3.5, the result is true for n=4⁠. Assume as induction hypothesis that the statement holds for Qi⁠, 4≤i≤n−1⁠. To complete the proof, one has to prove that if |Z6n|≤⌈n3⌉−1⁠, then Qn−Z6n is connected.

Case 1.

Assume that either V(Z6n)⊆V(Qn0) or V(Z6n)⊆V(Qn1)⁠. Without loss of generality, let us assume that V(Z6n)⊆V(Qn0)⁠. Note that Qn1 is connected and every vertex of Qn0−Z6n is connected to a vertex in Qn1 and hence Qn−Z6n is connected.

Case 2.

Suppose V(Z6n)∩V(Qn1)≠φ and V(Z6n)∩V(Qn0)≠φ⁠.

Case 2.1.

Assume that, for every 6-cycle X of Z6n⁠, V(X)⊂V(Qn1) or V(X)⊂V(Qn0)⁠. In this case, we have the number of 6-cycles of Z6n and Qn1 is at most ⌈n3⌉−2 and |V(Z6n)∩V(Qn0)|≤⌈n3⌉−2⁠. Note that ⌈n3⌉−1≤⌈n−13⌉ and so ⌈n3⌉−2<⌈n−13⌉⁠. By the induction hypothesis, κ(Qni;C6)≥⌈n−13⌉ and thus Qni−(Z6n)i is connected for i∈{0,1}⁠. Since 6(⌈n3⌉−2)<6(n+33−2)=2(n−3)<2(n−2)≤2n−2⁠, for i=0,1⁠, Qni−(Z6n)i contains more than 2n−12 vertices. Hence there exists a vertex u∈Qn1−(Z6n)1 which is adjacent to (u)n−1∈Qn0−(Z6n)0⁠. Hence Qn−Z6n is connected.

Case 2.2.

Assume that V(Z6n)∩V(Qn1)≠φ and V(Z6n)∩V(Qn0)≠φ and there is a 6-cycle X∈Z6n such that V(X)∩V(Qni)≠φ for each i=0,1⁠.

Let Z6n={X1,X2,…,Xm}⁠, m≤⌈n3⌉−1 where each Xi is a 6-cycle. For any Xi∈Z6n⁠, the elements of V(Xi) differ from one another in at most 3 coordinates. Let us name those coordinates as ki1,ki2,ki3⁠. i.e., if 0 (or 1) is the pth coordinate of an element in V(Xi) for p≠ki1,ki2,ki3⁠, then every element of V(Xi) has 0 (or 1⁠) as the pth coordinate.

Hence in total we have 3m such coordinates ki1,ki2,ki3⁠, for 1≤i≤m (not necessarily distinct) corresponding to all elements in Z6n⁠. Further, 3m≤3⌈n3⌉−3<3(n+33)−3=n⁠. Thus there exists k∈{1,2…,n} such that k∉{ki1,ki2,ki3} for each 1≤i≤m⁠. This means that kth coordinate of V(Xi) is same in all the 6 elements of Xi⁠. i.e., the kth coordinate of all the elements of V(Xi) are equal.

Let us partition the vertices of Qn into two subsets Vj={x=(x1…xn)∈V(Qn):xk=j,k is the index identified above},j∈{0,1}⁠. By the above arguments, for every i⁠, 1≤i≤m⁠, either V(Xi)⊆V0 or V(Xi)⊆V1⁠. Note that both the induced subgraphs 〈V0〉 and 〈V1〉 of Qn are isomorphic to Qn−1⁠. Now, if Z6n⊆Vj for some j∈{0,1}⁠, proceeding as in Case 1, Qn−Z6n is connected. Otherwise, proceeding as in Case 2.1, Qn−Z6n is connected. Thus, κ(Qn;C6)≥⌈n3⌉⁠. □

Figure 4 illustrates the C6-structure connectivity of Q4⁠. In Figure 4, the structure cut is indicated with the dotted lines.

Since Q3 is connected and by Lemma 3.3, κ(Q3;C6)≤⌈33⌉=1⁠, κ(Q3;C6)=1=⌈33⌉⁠. Also, by Lemma 3.3 and 3.6, we have the following result.

Theorem 3.7.

For n≥3, κs(Qn;C6)≤κ(Qn;C6)=⌈n3⌉.

This research work is supported through the INSPIRE, India programme (IF160672) of Department of Science and Technology, Government of India for the second author. 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. The publisher wishes to inform readers that the article “Structure and substructure connectivity of circulant graphs and hypercubes” was originally published by the previous publisher of the Arab Journal of Mathematical Sciences and the pagination of this article has been subsequently changed. There has been no change to the content of the article. This change was necessary for the journal to transition from the previous publisher to the new one. The publisher sincerely apologises for any inconvenience caused. To access and cite this article, please use Tamizh Chelvam, T., Sivagami, M., “Structure and substructure connectivity of circulant graphs and hypercubes”, Arab Journal of Mathematical Sciences, Vol. 27 No. 1, pp. 94-103. The original publication date for this paper was 24/10/2019.

[1]
N.
 
Biggs
,
Algebraic Graph Theory
,
Cambridge University Press
,
1993
.
[2]
G.
 
Chatr
,
P.
 
Zhang
,
Introduction To Graph Theory
,
Tata McGraw-Hill
,
2006
.
[3]
Italo J.
 
Dejter
,
Perfect domination in regular grid graphs
,
Australas. J. Combin.
 
42
(
2008
)
99
–
114
.
[4]
S.
 
Lakshmivarahan
,
S.K.
 
Dhall
,
Ring, torus, hypercube architectures algorithms for parallel computing
,
Parallel Comput
.
25
(
1999
)
1877
–
1906
.
[5]
J.
 
Lee
,
Independent perfect domination sets in Cayley graphs
,
J. Graph Theory
 
37
(
4
) (
2001
)
213
–
219
.
[6]
Cheng-Kuan
 
Lin
,
Lili
 
Zhang
,
Jianxi
 
Fan
,
Dajin
 
Wang
,
Structure connectivity and substructure connectivity of hypercubes
,
Theoret. Comput. Sci.
 
634
(
2016
)
97
–
107
, .
[7]
S.A.
 
Mane
,
Structure connectivity of hypercubes
,
AKCE Int. J. Graphs Comb.
 
15
(
1
) (
2018
)
49
–
52
.
[8]
T.
 
Tamizh Chelvam
,
I.
 
Rani
,
Independent domination number of Cayley graphs on ℤn
.,
J. Combin. Math. Combin. Comput.
 
69
(
2009
)
251
–
255
.
[9]
T.
 
Tamizh Chelvam
,
M.
 
Sivagami
,
Domination in Cayley graphs: A survey
,
AKCE Int. J. Graphs Comb.
 
16
(
2019
)
27
–
40
, .
[10]
J.F.
 
Wang
,
An investigation of the network reliability properties of circulant graphs (Doctoral dissertation)
,
Stevens Institute of Technology
,
1983
.
Published in the Arab Journal of Mathematical Sciences. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/legalcode

Data & Figures

Figure 1

κ(G)=2,κ(G;C5)=1,κ(G;C4)=3⁠.

Figure 1

κ(G)=2,κ(G;C5)=1,κ(G;C4)=3⁠.

Close Figure 1
Figure 2

K3-structure cut of Cir (10,{1,2,3,7,8,9})⁠.

Figure 2

K3-structure cut of Cir (10,{1,2,3,7,8,9})⁠.

Close Figure 2
Figure 3

K1,4-substructure cut of Cir(16,{1,2,14,15})⁠.

Figure 3

K1,4-substructure cut of Cir(16,{1,2,14,15})⁠.

Close Figure 3
Figure 4

C6-structure cut of Q4⁠.

Figure 4

C6-structure cut of Q4⁠.

Close Figure 4

Supplements

References

[1]
N.
 
Biggs
,
Algebraic Graph Theory
,
Cambridge University Press
,
1993
.
[2]
G.
 
Chatr
,
P.
 
Zhang
,
Introduction To Graph Theory
,
Tata McGraw-Hill
,
2006
.
[3]
Italo J.
 
Dejter
,
Perfect domination in regular grid graphs
,
Australas. J. Combin.
 
42
(
2008
)
99
–
114
.
[4]
S.
 
Lakshmivarahan
,
S.K.
 
Dhall
,
Ring, torus, hypercube architectures algorithms for parallel computing
,
Parallel Comput
.
25
(
1999
)
1877
–
1906
.
[5]
J.
 
Lee
,
Independent perfect domination sets in Cayley graphs
,
J. Graph Theory
 
37
(
4
) (
2001
)
213
–
219
.
[6]
Cheng-Kuan
 
Lin
,
Lili
 
Zhang
,
Jianxi
 
Fan
,
Dajin
 
Wang
,
Structure connectivity and substructure connectivity of hypercubes
,
Theoret. Comput. Sci.
 
634
(
2016
)
97
–
107
, .
[7]
S.A.
 
Mane
,
Structure connectivity of hypercubes
,
AKCE Int. J. Graphs Comb.
 
15
(
1
) (
2018
)
49
–
52
.
[8]
T.
 
Tamizh Chelvam
,
I.
 
Rani
,
Independent domination number of Cayley graphs on ℤn
.,
J. Combin. Math. Combin. Comput.
 
69
(
2009
)
251
–
255
.
[9]
T.
 
Tamizh Chelvam
,
M.
 
Sivagami
,
Domination in Cayley graphs: A survey
,
AKCE Int. J. Graphs Comb.
 
16
(
2019
)
27
–
40
, .
[10]
J.F.
 
Wang
,
An investigation of the network reliability properties of circulant graphs (Doctoral dissertation)
,
Stevens Institute of Technology
,
1983
.

Languages

or Create an Account

Close subscription notice
Close access options