Introduction
Let G G G be a simple connected graph with vertex set V ( G ) = { v i : 1 ≤ i ≤ p } V(G)=\{v_i:1\le i\le p\} V ( G ) = { v i : 1 ≤ i ≤ p } and edge set E ( G ) = { e j : 1 ≤ j ≤ q } E(G)=\{e_j:1\le j\le q\} E ( G ) = { e j : 1 ≤ j ≤ q } . The adjacency matrix of G G G , A ( G ) = ( a i j ) A(G)=(a_{ij}) A ( G ) = ( a ij ) is a p × p p\times p p × p symmetric matrix such that a i j = 1 a_{ij}=1 a ij = 1 , if v i v_i v i and v j v_j v j are adjacent, and 0 otherwise. Let λ 1 , λ 2 , … , λ p \lambda_1,\lambda_2,\ldots,\lambda_p λ 1 , λ 2 , … , λ p be the eigenvalues of A ( G ) A(G) A ( G ) (adjacency eigenvalues of G G G ). The adjacency spectrum of G G G , spec ( G ) \operatorname{spec}(G) spec ( G ) , is the collection of all eigenvalues of A ( G ) A(G) A ( G ) along with their multiplicities. Spectral graph theory provides a powerful framework for investigating the structural and combinatorial properties of graphs through the eigenvalues of matrices associated with them. Among the most commonly studied matrices are the adjacency matrix, the Laplacian matrix, and the distance matrix, each of which captures different aspects of a graph’s structure. In particular, the distance matrix has attracted considerable attention owing to its applications in communication networks, chemistry, and various optimization problems; see [1 ] .
The distance between any two vertices v i , v j ∈ V ( G ) v_i,v_j\in V(G) v i , v j ∈ V ( G ) , d ( v i , v j ) d(v_i,v_j) d ( v i , v j ) is defined as the length of the shortest path connecting them in G G G . In 1971, Graham and Pollack [5 ] introduced the distance matrix of a connected graph to study data communication problems. The distance matrix of G G G , D ( G ) D(G) D ( G ) is the p × p p\times p p × p symmetric matrix whose ( i , j ) (i,j) ( i , j ) -th entry is equal to d ( v i , v j ) d(v_i,v_j) d ( v i , v j ) . The eigenvalues of D ( G ) D(G) D ( G ) are called the D D D -eigenvalues of G G G and they along with their multiplicities form the D D D -spectrum of G G G . The D D D -eigenvalues of G G G have applications in various fields like data communication, chemistry, and molecular biology (see [2 ] and the references therein). Let η i \eta_i η i be the i t h i^{th} i t h largest eigenvalue of D ( G ) D(G) D ( G ) . The concept of distance energy of a graph was introduced in 2008 [8 ] . The distance energy (D D D -energy) of a graph G G G is defined as the sum of the absolute values of the eigenvalues of its distance matrix.
The subdivision graph S G S_G S G of a graph G G G [4 ] , is obtained from G G G by subdividing each edge of G G G using a new vertex, the vertex set of S G S_G S G is V ( G ) ∪ E ( G ) V(G) \cup E(G) V ( G ) ∪ E ( G ) . The concept of subdivision vertex join (respectively, subdivision edge join), which is obtained from S G 1 S_{G_1} S G 1 and G 2 G_2 G 2 by joining each vertex of V ( G 1 ) V(G_1) V ( G 1 ) (respectively, E ( G 1 ) E(G_1) E ( G 1 ) ) to every vertex of V ( G 2 ) V(G_2) V ( G 2 ) , was introduced in 2012 [7 ] . Later in 2024, Shamsher and Pirzada [12 ] generalized the notion of subdivision graphs as follows:
Let G G G be a graph with vertex set V ( G ) = { v 1 , v 2 , … , v p } V(G) = \{v_1,v_2,\ldots,v_p\} V ( G ) = { v 1 , v 2 , … , v p } and edge set E ( G ) = { e 1 , e 2 , … , e q } E(G) = \{e_1,e_2,\ldots,e_q\} E ( G ) = { e 1 , e 2 , … , e q } . The generalized subdivision graph of G G G , denoted by S G ( n , m ) S_G(n,m) S G ( n , m ) , is the graph obtained from S G S_G S G by replacing each vertex v i v_i v i , i = 1 , 2 , … , p i=1,2,\ldots,p i = 1 , 2 , … , p , with n n n new vertices having the same neighborhood as v i v_i v i , followed by replacing each edge e j e_j e j , j = 1 , 2 , … , q j=1,2,\ldots,q j = 1 , 2 , … , q , with m m m new vertices adjacent to the neighbors of e j e_j e j . Equivalently, the graph S G ( n , m ) S_G(n,m) S G ( n , m ) has vertex set
( V ( G ) × { 1 , 2 , … , n } ) ∪ ( E ( G ) × { 1 , 2 , … , m } ) , \left(V(G) \times\{1,2,\ldots,n\}\right) \cup\left(E(G) \times\{1,2,\ldots,m\}\right), ( V ( G ) × { 1 , 2 , … , n } ) ∪ ( E ( G ) × { 1 , 2 , … , m } ) ,
where two vertices ( v l , i ) (v_l,i) ( v l , i ) and ( e m , j ) (e_m,j) ( e m , j ) , are adjacent whenever the edge e m ∈ E ( G ) e_m \in E(G) e m ∈ E ( G ) is incident with the vertex v l ∈ V ( G ) v_l \in V(G) v l ∈ V ( G ) in G G G . Motivated by this generalized subdivision graph, we introduce the concept of generalized subdivision vertex join and generalized subdivision edge join of two graphs G 1 G_1 G 1 and G 2 G_2 G 2 as follows.
Definition 1.1. The generalized subdivision vertex join S G 1 ( n , m ) ∨ ˙ G 2 S_{G_1}(n,m)\dot{\vee}G_2 S G 1 ( n , m ) ∨ ˙ G 2 is the graph obtained from S G 1 ( n , m ) S_{G_1}(n,m) S G 1 ( n , m ) and G 2 G_2 G 2 by joining each vertex of V ( G 1 ) × { 1 , 2 , … , n } V(G_1)\times\{1,2,\ldots,n\} V ( G 1 ) × { 1 , 2 , … , n } to every vertex of V ( G 2 ) V(G_2) V ( G 2 ) .
Note that, if G i G_i G i is a graph on p i p_i p i vertices and q i q_i q i edges, i = 1 , 2 i=1,2 i = 1 , 2 , then S G 1 ( n , m ) ∨ ˙ G 2 S_{G_1}(n,m)\dot{\vee}G_2 S G 1 ( n , m ) ∨ ˙ G 2 has p 1 n + q 1 m + p 2 p_1n+q_1m+p_2 p 1 n + q 1 m + p 2 vertices and p 2 + p 1 n p 2 + ( p 1 + q 1 ) m n p_2+p_1np_2+(p_1+q_1)mn p 2 + p 1 n p 2 + ( p 1 + q 1 ) mn edges.
Example 1.2. Let G 1 = K 2 G_1=K_2 G 1 = K 2 and G 2 = C 3 G_2=C_3 G 2 = C 3 . Then, S K 2 ( 2 , 2 ) ∨ ˙ C 3 S_{K_2}(2,2)\dot{\vee}C_3 S K 2 ( 2 , 2 ) ∨ ˙ C 3 is depicted in Figure 1 .
Figure 1. S K 2 ( 2 , 2 ) ∨ ˙ K 2 S_{K_2}(2,2)\dot{\vee}K_2 S K 2 ( 2 , 2 ) ∨ ˙ K 2
Definition 1.3. The generalized subdivision edge join S G 1 ( n , m ) ∨ G 2 S_{G_1}(n,m)\vee G_2 S G 1 ( n , m ) ∨ G 2 is the graph obtained from S G 1 ( n , m ) S_{G_1}(n,m) S G 1 ( n , m ) and G 2 G_2 G 2 by joining each vertex of E ( G 1 ) × { 1 , 2 , … , m } E(G_1)\times\{1,2,\ldots,m\} E ( G 1 ) × { 1 , 2 , … , m } to every vertex of V ( G 2 ) V(G_2) V ( G 2 ) .
Note that, if G i G_i G i is a graph on p i p_i p i vertices and q i q_i q i edges, i = 1 , 2 i=1,2 i = 1 , 2 , then S G 1 ( n , m ) ∨ G 2 S_{G_1}(n,m)\vee G_2 S G 1 ( n , m ) ∨ G 2 has p 1 n + q 1 m + p 2 p_1n+q_1m+p_2 p 1 n + q 1 m + p 2 vertices and p 2 + q 1 m p 2 + ( p 1 + q 1 ) m n p_2+q_1mp_2+(p_1+q_1)mn p 2 + q 1 m p 2 + ( p 1 + q 1 ) mn edges.
Example 1.4. Let G 1 = K 2 G_1=K_2 G 1 = K 2 and G 2 = C 3 G_2=C_3 G 2 = C 3 . Then, S K 2 ( 2 , 2 ) ∨ C 3 S_{K_2}(2,2)\vee C_3 S K 2 ( 2 , 2 ) ∨ C 3 is depicted in Figure 2 .
Figure 2. S K 2 ( 2 , 2 ) ∨ K 2 S_{K_2}(2,2)\vee K_2 S K 2 ( 2 , 2 ) ∨ K 2
Two non-isomorphic graphs of the same order are cospectral (respectively, D D D -cospectral) if they have the same adjacency spectrum (respectively, D D D -spectrum). Non-isomorphic, non-cospectral (respectively, non D D D -cospectral) graphs of the same order are equienergetic (respectively, distance equienergetic) if their adjacency energies (respectively, distance energies) are the same. The construction of non-isomorphic equienergetic graphs with respect to various graph matrices is an important problem in spectral graph theory. In particular, several constructions of equienergetic and distance equienergetic graphs have been investigated in the literature; see [9 , 10 , 13 , 14 , 11 ] . In this work, we investigate the distance spectra of generalized subdivision vertex (respectively, edge) join of regular graphs. In particular, we derive new spectral results that encompass the main results of Indulal et al. (2019) [6 ] as special cases. In this work, we study the distance spectra of generalized subdivision vertex (respectively, edge) join of regular graphs. Our results generalize the main results of Indulal et al. (2019) [6 ] . Furthermore, using these results, we construct families of non-D D D -cospectral D D D -equienergetic graphs with diameter 4.
This article is organized as follows: Section 2 presents the preliminaries and notations used throughout. In Section 3, we present the distance spectra of the class of graphs obtained by the generalized subdivision vertex (respectively, edge) join of regular graphs. Finally, in Section 4, we construct families of D D D -equienergetic and D D D -cospectral graphs of diameter 4.
Preliminaries
The incidence matrix R = R ( G ) = ( r i j ) R=R(G)=(r_{ij}) R = R ( G ) = ( r ij ) of G G G , is the p × q p\times q p × q matrix such that r i j = 1 r_{ij}=1 r ij = 1 if and only if vertex v i v_i v i and edge e j e_j e j are incident in G G G . The line graph L ( G ) L(G) L ( G ) of G G G is the graph with its vertex set as the edge set of G G G , and two vertices in L ( G ) L(G) L ( G ) become adjacent if the corresponding edges of G G G share a common end vertex. The adjacency matrix of L ( G ) L(G) L ( G ) is denoted by B ( G ) B(G) B ( G ) . We indicate by K p K_p K p the complete graph, by C p C_p C p the cycle, each of order p p p . Throughout this paper, J p × q J_{p\times q} J p × q and I p × p I_{p\times p} I p × p represent the all-ones matrix and the identity matrix of the indicated orders, respectively. The Kronecker product A ⊗ B A\otimes B A ⊗ B , of two matrices A = ( a i j ) n × m A=(a_{ij})_{n\times m} A = ( a ij ) n × m and B = ( b i j ) p × q B=(b_{ij})_{p\times q} B = ( b ij ) p × q is obtained by replacing the ( i , j ) (i,j) ( i , j ) -th entry in A A A by a i j B a_{ij}B a ij B . For matrices A A A , B B B , C C C , and D D D of appropriate orders, we have ( A ⊗ B ) ( C ⊗ D ) = A C ⊗ B D (A\otimes B)(C\otimes D)=AC\otimes BD ( A ⊗ B ) ( C ⊗ D ) = A C ⊗ B D . Also, for convenience, we denote I p × p ⊗ ( J − I ) n × n I_{p\times p}\otimes(J-I)_{n\times n} I p × p ⊗ ( J − I ) n × n as I ~ p n × p n \widetilde{I}_{pn\times pn} I p n × p n . The equitable quotient matrix [3 ] of a partitioned matrix is denoted by M M M . The following lemmas provide the foundation for the results developed in the subsequent sections.
Lemma 2.1. [4 ] Let G G G be an r r r -regular, ( p , q ) (p,q) ( p , q ) graph with an adjacency matrix A A A , and the incidence matrix R R R . Let L ( G ) L(G) L ( G ) be the line graph of G G G . Then, R R T = A + r I RR^{T}=A+rI R R T = A + r I , R T R = B + 2 I R^{T}R=B+2I R T R = B + 2 I , where B B B is the adjacency matrix of L ( G ) L(G) L ( G ) , also J R = 2 J = R T J JR=2J=R^{T}J J R = 2 J = R T J and J R T = r J = R J JR^{T}=rJ=RJ J R T = r J = R J .
Lemma 2.2. [4 ] Let G G G be an r r r -regular, ( p , q ) (p,q) ( p , q ) graph with spec ( G ) = { r , λ 2 , … , λ p } \operatorname{spec}(G)=\{r,\lambda_2,\ldots,\lambda_p\} spec ( G ) = { r , λ 2 , … , λ p } . Then,
spec ( L ( G ) ) = ( 2 r − 2 λ 2 + r − 2 ⋯ λ p + r − 2 − 2 1 1 ⋯ 1 q − p ) \operatorname{spec}(L(G))=
\begin{pmatrix}
2r-2 & \lambda_2+r-2 & \cdots& \lambda_p+r-2 & -2 \\
1 & 1 & \cdots& 1 & q-p
\end{pmatrix} spec ( L ( G )) = ( 2 r − 2 1 λ 2 + r − 2 1 ⋯ ⋯ λ p + r − 2 1 − 2 q − p )
Also, Z Z Z is an eigenvector corresponding to the eigenvalue − 2 -2 − 2 if and only if R Z = 0 RZ=0 R Z = 0 .
Distance Spectra of Generalized Subdivision Vertex (Edge) Joins
In this section, we determine the distance spectra of the generalized subdivision vertex, and edge join of regular graphs G 1 G_1 G 1 and G 2 G_2 G 2 in terms of their adjacency spectra and the adjacency spectrum of the line graph of G 1 G_1 G 1 .
The following lemma establishes the distance between vertices in S G 1 ( n , n ) ∨ ˙ G 2 SG_1(n,n)\dot{\vee}G_2 S G 1 ( n , n ) ∨ ˙ G 2 .
Proposition 3.1. Let V ( G ) = { v 1 , v 2 , … , v p } V(G)=\{v_1,v_2,\ldots,v_p\} V ( G ) = { v 1 , v 2 , … , v p } and E ( G ) = { e 1 , e 2 , … , e q } E(G)=\{e_1,e_2,\ldots,e_q\} E ( G ) = { e 1 , e 2 , … , e q } . Then, the distance between any two distinct vertices, say x x x and y y y , in S G 1 ( n , n ) ∨ ˙ G 2 SG_1(n,n)\dot{\vee}G_2 S G 1 ( n , n ) ∨ ˙ G 2 is as follows:
d ( x , y ) = { 2 if x = ( v i , j ) , y = ( v l , s ) , 1 if x = ( v i , j ) , y = ( e l , s ) and v i is incident with e l in G 1 , 3 if x = ( v i , j ) , y = ( e l , s ) and v i is not incident with e l in G 1 , 1 if x = ( v i , j ) and y ∈ V ( G 2 ) , 2 if x = ( e i , j ) , y = ( e l , s ) and e i , e l have a common end point in G 1 , 4 if x = ( e i , j ) , y = ( e l , s ) and e i , e l have no common end point in G 1 , 2 if x = ( e i , j ) and y ∈ V ( G 2 ) , 1 if x , y ∈ V ( G 2 ) and x , y are adjacent in V ( G 2 ) , 2 if x , y ∈ V ( G 2 ) and x , y are not adjacent in V ( G 2 ) . d(x,y)=
\begin{cases}
2 & \text{if }x=(v_i,j),y=(v_l,s),\\
1 & \text{if }x=(v_i,j),y=(e_l,s)\text{ and }v_i\text{ is incident with }e_l\text{ in }G_1,\\
3 & \text{if }x=(v_i,j),y=(e_l,s)\text{ and }v_i\text{ is not incident with }e_l\text{ in }G_1,\\
1 & \text{if }x=(v_i,j)\text{ and }y\in V(G_2),\\
2 & \text{if }x=(e_i,j),y=(e_l,s)\text{ and }e_i,e_l\text{ have a common end point in }G_1,\\
4 & \text{if }x=(e_i,j),y=(e_l,s)\text{ and }e_i,e_l\text{ have no common end point in }G_1,\\
2 & \text{if }x=(e_i,j)\text{ and }y\in V(G_2),\\
1 & \text{if }x,y\in V(G_2)\text{ and }x,y\text{ are adjacent in }V(G_2),\\
2 & \text{if }x,y\in V(G_2)\text{ and }x,y\text{ are not adjacent in }V(G_2).
\end{cases} d ( x , y ) = ⎩ ⎨ ⎧ 2 1 3 1 2 4 2 1 2 if x = ( v i , j ) , y = ( v l , s ) , if x = ( v i , j ) , y = ( e l , s ) and v i is incident with e l in G 1 , if x = ( v i , j ) , y = ( e l , s ) and v i is not incident with e l in G 1 , if x = ( v i , j ) and y ∈ V ( G 2 ) , if x = ( e i , j ) , y = ( e l , s ) and e i , e l have a common end point in G 1 , if x = ( e i , j ) , y = ( e l , s ) and e i , e l have no common end point in G 1 , if x = ( e i , j ) and y ∈ V ( G 2 ) , if x , y ∈ V ( G 2 ) and x , y are adjacent in V ( G 2 ) , if x , y ∈ V ( G 2 ) and x , y are not adjacent in V ( G 2 ) .
Lemma 3.2. Let G i G_i G i be an r i r_i r i -regular graph on p i p_i p i vertices and q i q_i q i edges with the adjacency matrix A ( G i ) A(G_i) A ( G i ) and adjacency spectrum { r i , λ i 2 , λ i 3 , … , λ i p i } \{r_i,\lambda_{i2},\lambda_{i3},\ldots,\lambda_{ip_i}\} { r i , λ i 2 , λ i 3 , … , λ i p i } , i = 1 , 2 i=1,2 i = 1 , 2 . Then, the distance spectrum of S G 1 ( n , n ) ∨ ˙ G 2 SG_1(n,n)\dot{\vee}G_2 S G 1 ( n , n ) ∨ ˙ G 2 consists of − 2 -2 − 2 with multiplicity ( n − 1 ) ( p 1 + q 1 ) (n-1)(p_1+q_1) ( n − 1 ) ( p 1 + q 1 ) ; 2 ( n − 1 ) 2(n-1) 2 ( n − 1 ) with multiplicity q 1 − 1 q_1-1 q 1 − 1 ; − 2 ( n λ 1 j + n r 1 + 1 ) -2(n\lambda_{1j}+nr_1+1) − 2 ( n λ 1 j + n r 1 + 1 ) , j = 2 , 3 , … , p 1 j=2,3,\ldots,p_1 j = 2 , 3 , … , p 1 ; − 2 − λ 2 m -2-\lambda_{2m} − 2 − λ 2 m , m = 2 , 3 , … , p 2 m=2,3,\ldots,p_2 m = 2 , 3 , … , p 2 ; and the eigenvalues of the matrix,
( 2 ( n p 1 − 1 ) n ( 3 q 1 − 2 r 1 ) p 2 n ( 3 p 1 − 4 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 p 2 p 1 n 2 q 1 n 2 ( p 2 − 1 ) − r 2 ) . \begin{pmatrix}
2(np_1-1) & n(3q_1-2r_1) & p_2\\
n(3p_1-4) & 4(q_1-r_1)n+2(n-1) & 2p_2\\
p_1n & 2q_1n & 2(p_2-1)-r_2
\end{pmatrix}. 2 ( n p 1 − 1 ) n ( 3 p 1 − 4 ) p 1 n n ( 3 q 1 − 2 r 1 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 q 1 n p 2 2 p 2 2 ( p 2 − 1 ) − r 2 . Proof. By suitable labeling of the vertices of S G 1 ( n , n ) ∨ ˙ G 2 SG_1(n,n)\dot{\vee}G_2 S G 1 ( n , n ) ∨ ˙ G 2 , the distance matrix D ( S G 1 ( n , n ) ∨ ˙ G 2 ) D(SG_1(n,n)\dot{\vee}G_2) D ( S G 1 ( n , n ) ∨ ˙ G 2 ) can be represented as
( ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) + 2 I ~ p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n J p 1 n × p 2 ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) + 2 I ~ q 1 n × q 1 n 2 J q 1 n × p 2 J p 2 × p 1 n 2 J p 2 × q 1 n ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 ) . \begin{pmatrix}
(2(J-I)_{p_1\times p_1}\otimes J_{n\times n})+2\widetilde{I}_{p_1n\times p_1n} &
(3J-2R(G_1))_{p_1\times q_1}\otimes J_{n\times n} &
J_{p_1n\times p_2}\\
(3J-2R(G_1)^{T})_{q_1\times p_1}\otimes J_{n\times n} &
((4(J-I)-2B(G_1))_{q_1\times q_1}\otimes J_{n\times n})+2\widetilde{I}_{q_1n\times q_1n} &
2J_{q_1n\times p_2}\\
J_{p_2\times p_1n} & 2J_{p_2\times q_1n} & (2(J-I)-A(G_2))_{p_2\times p_2}
\end{pmatrix}. ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) + 2 I p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n J p 2 × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n (( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) + 2 I q 1 n × q 1 n 2 J p 2 × q 1 n J p 1 n × p 2 2 J q 1 n × p 2 ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 . For s = 1 , 2 , 3 , … , p 1 s=1,2,3,\ldots,p_1 s = 1 , 2 , 3 , … , p 1 , consider the vector
z ( s , t ) = ( x 1 , 1 , x 1 , 2 , x 1 , 3 , … , x 1 , n , x 2 , 1 , x 2 , 2 , x 2 , 3 , … , x 2 , n , … , x p 1 , 1 , x p 1 , 2 , x p 1 , 3 , … , x p 1 , n , z^{(s,t)}=(x_{1,1},x_{1,2},x_{1,3},\ldots,x_{1,n},x_{2,1},x_{2,2},x_{2,3},\ldots,x_{2,n},\ldots,x_{p_1,1},x_{p_1,2},x_{p_1,3},\ldots,x_{p_1,n}, z ( s , t ) = ( x 1 , 1 , x 1 , 2 , x 1 , 3 , … , x 1 , n , x 2 , 1 , x 2 , 2 , x 2 , 3 , … , x 2 , n , … , x p 1 , 1 , x p 1 , 2 , x p 1 , 3 , … , x p 1 , n , y 1 , 1 , y 1 , 2 , y 1 , 3 , … , y 1 , n , y 2 , 1 , y 2 , 2 , y 2 , 3 , … , y 2 , n , … , y q 1 , 1 , y q 1 , 2 , y q 1 , 3 , … , y q 1 , n , y_{1,1},y_{1,2},y_{1,3},\ldots,y_{1,n},y_{2,1},y_{2,2},y_{2,3},\ldots,y_{2,n},\ldots,y_{q_1,1},y_{q_1,2},y_{q_1,3},\ldots,y_{q_1,n}, y 1 , 1 , y 1 , 2 , y 1 , 3 , … , y 1 , n , y 2 , 1 , y 2 , 2 , y 2 , 3 , … , y 2 , n , … , y q 1 , 1 , y q 1 , 2 , y q 1 , 3 , … , y q 1 , n , w 1 , 1 , w 1 , 2 , w 1 , 3 , … , w 1 , p 2 ) ∈ R N , (1) w_{1,1},w_{1,2},w_{1,3},\ldots,w_{1,p_2})\in\mathbb{R}^{N}, \tag*{(1)} w 1 , 1 , w 1 , 2 , w 1 , 3 , … , w 1 , p 2 ) ∈ R N , (1) where N = p 1 n + q 1 n + p 2 N=p_1n+q_1n+p_2 N = p 1 n + q 1 n + p 2 , and
x s , 1 = 1 and x s , t = − 1 , for t = 2 , 3 , … , n x_{s,1}=1\text{ and }x_{s,t}=-1,\text{ for }t=2,3,\ldots,n x s , 1 = 1 and x s , t = − 1 , for t = 2 , 3 , … , n and all remaining coordinates of z ( s , t ) z^{(s,t)} z ( s , t ) are equal to 0. Then,
D ( S G 1 ( n , n ) ∨ ˙ G 2 ) z ( s , t ) = − 2 z ( s , t ) , s = 1 , 2 , 3 , … , p 1 . D(SG_1(n,n)\dot{\vee}G_2)z^{(s,t)}=-2z^{(s,t)},\quad s=1,2,3,\ldots,p_1. D ( S G 1 ( n , n ) ∨ ˙ G 2 ) z ( s , t ) = − 2 z ( s , t ) , s = 1 , 2 , 3 , … , p 1 .
This construction yields ( n − 1 ) p 1 (n-1)p_1 ( n − 1 ) p 1 linearly independent eigenvectors of D ( S G 1 ( n , n ) ∨ ˙ G 2 ) D(SG_1(n,n)\dot\vee G_2) D ( S G 1 ( n , n ) ∨ ˙ G 2 ) associated with the eigenvalue − 2 -2 − 2 .
Using similar argument, for k = 1 , 2 , 3 , … , q 1 k=1,2,3,\ldots,q_1 k = 1 , 2 , 3 , … , q 1 , consider the vector
h ( k , l ) = ( u 1 , 1 , u 1 , 2 , u 1 , 3 , … , u 1 , n , u 2 , 1 , u 2 , 2 , u 2 , 3 , … , u 2 , n , … , u p 1 , 1 , u p 1 , 2 , u p 1 , 3 , … , u p 1 , n , v 1 , 1 , v 1 , 2 , v 1 , 3 , … , v 1 , n , v 2 , 1 , v 2 , 2 , v 2 , 3 , … , v 2 , n , … , v q 1 , 1 , v q 1 , 2 , v q 1 , 3 , … , v q 1 , n , d 1 , 1 , d 1 , 2 , d 1 , 3 , … , d 1 , p 2 ) ∈ R N (2) h^{(k,l)}=\left(u_{1,1},u_{1,2},u_{1,3},\ldots,u_{1,n},u_{2,1},u_{2,2},u_{2,3},\ldots,u_{2,n},\ldots,u_{p_1,1},u_{p_1,2},u_{p_1,3},\ldots,u_{p_1,n},
v_{1,1},v_{1,2},v_{1,3},\ldots,v_{1,n},v_{2,1},v_{2,2},v_{2,3},\ldots,v_{2,n},\ldots,v_{q_1,1},v_{q_1,2},v_{q_1,3},\ldots,v_{q_1,n},
d_{1,1},d_{1,2},d_{1,3},\ldots,d_{1,p_2}\right)\in\mathbb{R}^{N} \tag*{(2)} h ( k , l ) = ( u 1 , 1 , u 1 , 2 , u 1 , 3 , … , u 1 , n , u 2 , 1 , u 2 , 2 , u 2 , 3 , … , u 2 , n , … , u p 1 , 1 , u p 1 , 2 , u p 1 , 3 , … , u p 1 , n , v 1 , 1 , v 1 , 2 , v 1 , 3 , … , v 1 , n , v 2 , 1 , v 2 , 2 , v 2 , 3 , … , v 2 , n , … , v q 1 , 1 , v q 1 , 2 , v q 1 , 3 , … , v q 1 , n , d 1 , 1 , d 1 , 2 , d 1 , 3 , … , d 1 , p 2 ) ∈ R N (2)
where
v k , 1 = 1 , and v k , l = − 1 , for l = 2 , 3 , … , n , v_{k,1}=1,\quad\text{and}\quad v_{k,l}=-1,\quad\text{for }l=2,3,\ldots,n, v k , 1 = 1 , and v k , l = − 1 , for l = 2 , 3 , … , n ,
and the remaining coordinates are all zero. Then,
D ( S G 1 ( n , n ) ∨ ˙ G 2 ) h ( k , l ) = − 2 h ( k , l ) , k = 1 , 2 , 3 , … , q 1 , D(SG_1(n,n)\dot\vee G_2)h^{(k,l)}=-2h^{(k,l)},\quad k=1,2,3,\ldots,q_1, D ( S G 1 ( n , n ) ∨ ˙ G 2 ) h ( k , l ) = − 2 h ( k , l ) , k = 1 , 2 , 3 , … , q 1 ,
which yields a further collection of q 1 ( n − 1 ) q_1(n-1) q 1 ( n − 1 ) linearly independent eigenvectors corresponding to the eigenvalue − 2 -2 − 2 . As the vectors z ( s , t ) z^{(s,t)} z ( s , t ) and h ( k , l ) h^{(k,l)} h ( k , l ) have disjoint supports, the two families together form a linearly independent set. Hence, − 2 -2 − 2 is an eigenvalue of D ( S G 1 ( n , n ) ∨ ˙ G 2 ) D(SG_1(n,n)\dot\vee G_2) D ( S G 1 ( n , n ) ∨ ˙ G 2 ) with multiplicity at least ( n − 1 ) ( p 1 + q 1 ) (n-1)(p_1+q_1) ( n − 1 ) ( p 1 + q 1 ) .
Let X ( j ) X^{(j)} X ( j ) be an eigenvector of A ( G 1 ) A(G_1) A ( G 1 ) corresponding to the eigenvalue λ 1 j ≠ r 1 \lambda_{1j}\ne r_1 λ 1 j = r 1 . Then, by the Perron-Frobenius theorem, X ( j ) X^{(j)} X ( j ) is orthogonal to J J J . By Lemma 2.1 , we have
R ( G 1 ) R ( G 1 ) T = A ( G 1 ) + r 1 I . R(G_1)R(G_1)^{T}=A(G_1)+r_1I. R ( G 1 ) R ( G 1 ) T = A ( G 1 ) + r 1 I .
R ( G 1 ) R ( G 1 ) T X ( j ) = ( A ( G 1 ) + r 1 I ) X ( j ) = ( λ 1 j + r 1 ) X ( j ) . \begin{aligned}
R(G_1)R(G_1)^{T}X^{(j)}=(A(G_1)+r_1I)X^{(j)} \\
&=(\lambda_{1j}+r_1)X^{(j)}.
\end{aligned} R ( G 1 ) R ( G 1 ) T X ( j ) = ( A ( G 1 ) + r 1 I ) X ( j ) = ( λ 1 j + r 1 ) X ( j ) .
Also, B ( G 1 ) = R ( G 1 ) T R ( G 1 ) − 2 I B(G_1)=R(G_1)^{T}R(G_1)-2I B ( G 1 ) = R ( G 1 ) T R ( G 1 ) − 2 I .
B ( G 1 ) R ( G 1 ) T X ( j ) = ( R ( G 1 ) T R ( G 1 ) − 2 I ) R ( G 1 ) T X ( j ) = R ( G 1 ) T ( A ( G 1 ) + r 1 I ) X ( j ) − 2 R ( G 1 ) T X ( j ) = ( λ 1 j + r 1 − 2 ) R ( G 1 ) T X ( j ) . \begin{aligned}
B(G_1)R(G_1)^{T}X^{(j)}=(R(G_1)^{T}R(G_1)-2I)R(G_1)^{T}X^{(j)} \\
&=R(G_1)^{T}(A(G_1)+r_1I)X^{(j)}-2R(G_1)^{T}X^{(j)} \\
&=(\lambda_{1j}+r_1-2)R(G_1)^{T}X^{(j)}.
\end{aligned} B ( G 1 ) R ( G 1 ) T X ( j ) = ( R ( G 1 ) T R ( G 1 ) − 2 I ) R ( G 1 ) T X ( j ) = R ( G 1 ) T ( A ( G 1 ) + r 1 I ) X ( j ) − 2 R ( G 1 ) T X ( j ) = ( λ 1 j + r 1 − 2 ) R ( G 1 ) T X ( j ) .
Therefore, R ( G 1 ) T X ( j ) R(G_1)^{T}X^{(j)} R ( G 1 ) T X ( j ) is an eigenvector of B ( G 1 ) B(G_1) B ( G 1 ) with an eigenvalue λ 1 j + r 1 − 2 \lambda_{1j}+r_1-2 λ 1 j + r 1 − 2 , which is different from its regularity as λ 1 j ≠ r 1 \lambda_{1j}\ne r_1 λ 1 j = r 1 . Then, again by the Perron-Frobenius theorem, R ( G 1 ) X ( j ) R(G_1)X^{(j)} R ( G 1 ) X ( j ) is orthogonal to J J J . Consider
ψ j = ( X p 1 × 1 ( j ) ⊗ J n × 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 O p 2 × 1 ) . \psi_j=
\begin{pmatrix}
X^{(j)}_{p_1\times1}\otimes J_{n\times1}\\
(R(G_1)^{T}X^{(j)})_{q_1\times1}\otimes J_{n\times1}\\
O_{p_2\times1}
\end{pmatrix}. ψ j = X p 1 × 1 ( j ) ⊗ J n × 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 O p 2 × 1 .
Then,
D ( S G 1 ( n , n ) ∨ ˙ G 2 ) ψ j = ( ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 I ~ p 1 n × p 1 n ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ( ( 3 J − 2 R ( G 1 ) ) q 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J n × n J n × 1 ) + ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) + 2 I ~ q 1 n × q 1 n ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ( J p 2 × p 1 ⊗ J 1 × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 ( J p 2 × q 1 ⊗ J 1 × n ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ) . \begin{aligned}
D(SG_1(n,n)\dot\vee G_2)\psi_j
&=
\begin{pmatrix}
(2(J-I)_{p_1\times p_1}\otimes J_{n\times n})(X^{(j)}_{p_1\times1}\otimes J_{n\times1})+
2\widetilde I_{p_1n\times p_1n}(X^{(j)}_{p_1\times1}\otimes J_{n\times1})+
((3J-2R(G_1))_{p_1\times q_1}\otimes J_{n\times n})((R(G_1)^TX^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
((3J-2R(G_1))_{q_1\times p_1}X^{(j)}_{p_1\times1})\otimes(J_{n\times n}J_{n\times1})+
((4(J-I)-2B(G_1))_{q_1\times q_1}\otimes J_{n\times n})((R(G_1)^TX^{(j)})_{q_1\times1}\otimes J_{n\times1})+
2\widetilde I_{q_1n\times q_1n}((R(G_1)^TX^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
(J_{p_2\times p_1}\otimes J_{1\times n})(X^{(j)}_{p_1\times1}\otimes J_{n\times1})+
2(J_{p_2\times q_1}\otimes J_{1\times n})((R(G_1)^TX^{(j)})_{q_1\times1}\otimes J_{n\times1})
\end{pmatrix}.
\end{aligned} D ( S G 1 ( n , n ) ∨ ˙ G 2 ) ψ j = ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 I p 1 n × p 1 n ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + (( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) (( 3 J − 2 R ( G 1 ) ) q 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J n × n J n × 1 ) + (( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) + 2 I q 1 n × q 1 n (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ( J p 2 × p 1 ⊗ J 1 × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 ( J p 2 × q 1 ⊗ J 1 × n ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) .
( ( 2 ( J − I ) p 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J n × n J n × 1 ) + ( ( 3 J − 2 R ) p 1 × q 1 ( R T X ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + 2 I ~ p 1 n × p 1 n ( X p 1 × 1 ( j ) ⊗ J n × 1 ) ( ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( n J n × 1 ) + ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) ( ( R T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ( J p 2 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J 1 × n J n × 1 ) + ( J p 2 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) ) = ( ( − 2 X p 1 × 1 ( j ) ) ⊗ ( n J n × 1 ) + ( − 2 R ( G 1 ) R ( G 1 ) T ) p 1 × 1 ⊗ ( n J n × 1 + ( I p 1 × p 1 ⊗ 2 ( J − I ) n × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) ( ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 X p 1 × 1 ( j ) + ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ 2 ( n − 1 ) J n × 1 ) ( O p 2 × 1 ⊗ n ) + 2 ( O p 2 × 1 ⊗ n ) ) = ( ( − 2 X p 1 × 1 ( j ) − 2 ( R ( G 1 ) R ( G 1 ) T X ( j ) ) p 1 × 1 ) ⊗ ( n J n × 1 ) + ( X p 1 × 1 ( j ) ⊗ 2 ( J − I ) n × n J n × 1 ) ( − 2 ( R ( G 1 ) T X ( j ) ) q 1 × 1 − 4 ( R ( G 1 ) T X ( j ) ) q 1 × 1 − 2 ( λ 1 j + r 1 − 2 ) ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ n J n × 1 + 2 ( n − 1 ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 ) = ( ( ( − 2 X p 1 × 1 ( j ) − 2 ( A ( G 1 ) + r 1 I ) X ( j ) ) p 1 × 1 ) ⊗ ( n J n × 1 ) + 2 ( n − 1 ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) n ( − 2 − 2 λ 1 j − 2 r 1 ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 ) = ( ( − 2 − 2 λ 1 j − 2 r 1 ) ( X p 1 × 1 ( j ) ⊗ n J n × 1 ) + 2 ( n − 1 ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) − 2 ( n λ 1 j + n r 1 + 1 ) ( ( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 ) = − 2 ( n λ 1 j + n r 1 + 1 ) ψ j . \begin{aligned}
&\begin{pmatrix}
(2(J-I)_{p_1\times p_1}X^{(j)}_{p_1\times1})\otimes(J_{n\times n}J_{n\times1})+\\
((3J-2R)_{p_1\times q_1}(R^T X^{(j)})_{q_1\times1})\otimes(J_{n\times n}J_{n\times1})+\\
2\widetilde I_{p_1n\times p_1n}(X^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
((3J-2R(G_1)^T)_{q_1\times p_1}X^{(j)}_{p_1\times1})\otimes(nJ_{n\times1})+\\
((4(J-I)-2B(G_1))_{q_1\times q_1}(R(G_1)^T X^{(j)})_{q_1\times1})\otimes(nJ_{n\times1})+\\
(I_{q_1\times q_1}\otimes2(J-I)_{n\times n})((R^T X^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
(J_{p_2\times p_1}X^{(j)}_{p_1\times1})\otimes(J_{1\times n}J_{n\times1})+\\
(J_{p_2\times q_1}(R(G_1)^T X^{(j)})_{q_1\times1})\otimes(J_{1\times n}J_{n\times1})
\end{pmatrix}\\
={}&\begin{pmatrix}
(-2X^{(j)}_{p_1\times1})\otimes(nJ_{n\times1})+(-2R(G_1)R(G_1)^T)_{p_1\times1}\otimes(nJ_{n\times1}+\\
(I_{p_1\times p_1}\otimes2(J-I)_{n\times n})(X^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
((3J-2R(G_1)^T)_{q_1\times p_1}X^{(j)}_{p_1\times1}+\\
(4(J-I)-2B(G_1))_{q_1\times q_1}(R(G_1)^T X^{(j)})_{q_1\times1})\otimes(nJ_{n\times1})+\\
((R(G_1)^T X^{(j)})_{q_1\times1}\otimes2(n-1)J_{n\times1})\\
(O_{p_2\times1}\otimes n)+2(O_{p_2\times1}\otimes n)
\end{pmatrix}\\
={}&\begin{pmatrix}
(-2X^{(j)}_{p_1\times1}-2(R(G_1)R(G_1)^T X^{(j)})_{p_1\times1})\otimes(nJ_{n\times1})+(X^{(j)}_{p_1\times1}\otimes2(J-I)_{n\times n}J_{n\times1})\\
(-2(R(G_1)^T X^{(j)})_{q_1\times1}-4(R(G_1)^T X^{(j)})_{q_1\times1}-2(\lambda_{1j}+r_1-2)(R(G_1)^T X^{(j)})_{q_1\times1})\otimes nJ_{n\times1}+\\
2(n-1)((R(G_1)^T X^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
O_{p_2\times1}
\end{pmatrix}\\
={}&\begin{pmatrix}
((-2X^{(j)}_{p_1\times1}-2(A(G_1)+r_1I)X^{(j)})_{p_1\times1})\otimes(nJ_{n\times1})+2(n-1)(X^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
n(-2-2\lambda_{1j}-2r_1)((R(G_1)^T X^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
O_{p_2\times1}
\end{pmatrix}\\
={}&\begin{pmatrix}
(-2-2\lambda_{1j}-2r_1)(X^{(j)}_{p_1\times1}\otimes nJ_{n\times1})+2(n-1)(X^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
-2(n\lambda_{1j}+nr_1+1)((R(G_1)^T X^{(j)})_{q_1\times1}\otimes J_{n\times1})\\
O_{p_2\times1}
\end{pmatrix}\\
={}&-2(n\lambda_{1j}+nr_1+1)\psi_j.
\end{aligned} = = = = = ( 2 ( J − I ) p 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J n × n J n × 1 ) + (( 3 J − 2 R ) p 1 × q 1 ( R T X ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + 2 I p 1 n × p 1 n ( X p 1 × 1 ( j ) ⊗ J n × 1 ) (( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 X p 1 × 1 ( j ) ) ⊗ ( n J n × 1 ) + (( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) (( R T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) ( J p 2 × p 1 X p 1 × 1 ( j ) ) ⊗ ( J 1 × n J n × 1 ) + ( J p 2 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) ( − 2 X p 1 × 1 ( j ) ) ⊗ ( n J n × 1 ) + ( − 2 R ( G 1 ) R ( G 1 ) T ) p 1 × 1 ⊗ ( n J n × 1 + ( I p 1 × p 1 ⊗ 2 ( J − I ) n × n ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) (( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 X p 1 × 1 ( j ) + ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ 2 ( n − 1 ) J n × 1 ) ( O p 2 × 1 ⊗ n ) + 2 ( O p 2 × 1 ⊗ n ) ( − 2 X p 1 × 1 ( j ) − 2 ( R ( G 1 ) R ( G 1 ) T X ( j ) ) p 1 × 1 ) ⊗ ( n J n × 1 ) + ( X p 1 × 1 ( j ) ⊗ 2 ( J − I ) n × n J n × 1 ) ( − 2 ( R ( G 1 ) T X ( j ) ) q 1 × 1 − 4 ( R ( G 1 ) T X ( j ) ) q 1 × 1 − 2 ( λ 1 j + r 1 − 2 ) ( R ( G 1 ) T X ( j ) ) q 1 × 1 ) ⊗ n J n × 1 + 2 ( n − 1 ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 (( − 2 X p 1 × 1 ( j ) − 2 ( A ( G 1 ) + r 1 I ) X ( j ) ) p 1 × 1 ) ⊗ ( n J n × 1 ) + 2 ( n − 1 ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) n ( − 2 − 2 λ 1 j − 2 r 1 ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 ( − 2 − 2 λ 1 j − 2 r 1 ) ( X p 1 × 1 ( j ) ⊗ n J n × 1 ) + 2 ( n − 1 ) ( X p 1 × 1 ( j ) ⊗ J n × 1 ) − 2 ( n λ 1 j + n r 1 + 1 ) (( R ( G 1 ) T X ( j ) ) q 1 × 1 ⊗ J n × 1 ) O p 2 × 1 − 2 ( n λ 1 j + n r 1 + 1 ) ψ j .
Thus, ψ j \psi_j ψ j is an eigenvector of D ( S G 1 ( n , n ) ∨ G ˙ 2 ) D(S_{G_1}(n,n)\vee\dot G_2) D ( S G 1 ( n , n ) ∨ G ˙ 2 ) with an eigenvalue − 2 ( n λ 1 j + n r 1 + 1 ) -2(n\lambda_{1j}+nr_1+1) − 2 ( n λ 1 j + n r 1 + 1 ) .
Now, − 2 -2 − 2 is an eigenvalue of B ( G 1 ) B(G_1) B ( G 1 ) with multiplicity q 1 − p 1 q_1-p_1 q 1 − p 1 . Let Z ( i ) Z^{(i)} Z ( i ) be an eigenvector of B ( G 1 ) B(G_1) B ( G 1 ) corresponding to the eigenvalue − 2 -2 − 2 . Then, Z ( i ) Z^{(i)} Z ( i ) is orthogonal to J J J , and by Lemma 2.2, R ( G 1 ) Z ( i ) = 0 R(G_1)Z^{(i)}=0 R ( G 1 ) Z ( i ) = 0 . Now, consider the vector χ i = ( O p 1 × 1 ⊗ J n × 1 Z q 1 × 1 ( i ) ⊗ J n × 1 O p 2 × 1 ) \chi_i=\begin{pmatrix}O_{p_1\times1}\otimes J_{n\times1}\\ Z^{(i)}_{q_1\times1}\otimes J_{n\times1}\\ O_{p_2\times1}\end{pmatrix} χ i = O p 1 × 1 ⊗ J n × 1 Z q 1 × 1 ( i ) ⊗ J n × 1 O p 2 × 1 . Then,
D ( S G 1 ( n , n ) ∨ G ˙ 2 ) χ i = ( ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 Z q 1 × 1 ( i ) ) ⊗ J n × n J n × 1 ( ( 4 ( J − I ) − 2 B ) q 1 × q 1 ⊗ J n × n ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) + 2 I ~ q 1 n × q 1 n ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) ( 2 J p 2 × q 1 ⊗ J 1 × n ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) ) D(S_{G_1}(n,n)\vee\dot G_2)\chi_i=
\begin{pmatrix}
((3J-2R(G_1))_{p_1\times q_1}Z^{(i)}_{q_1\times1})\otimes J_{n\times n}J_{n\times1}\\
((4(J-I)-2B)_{q_1\times q_1}\otimes J_{n\times n})(Z^{(i)}_{q_1\times1}\otimes J_{n\times1})+2\widetilde I_{q_1n\times q_1n}(Z^{(i)}_{q_1\times1}\otimes J_{n\times1})\\
(2J_{p_2\times q_1}\otimes J_{1\times n})(Z^{(i)}_{q_1\times1}\otimes J_{n\times1})
\end{pmatrix} D ( S G 1 ( n , n ) ∨ G ˙ 2 ) χ i = (( 3 J − 2 R ( G 1 ) ) p 1 × q 1 Z q 1 × 1 ( i ) ) ⊗ J n × n J n × 1 (( 4 ( J − I ) − 2 B ) q 1 × q 1 ⊗ J n × n ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) + 2 I q 1 n × q 1 n ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) ( 2 J p 2 × q 1 ⊗ J 1 × n ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 )
( O p 1 × 1 ⊗ n J n × 1 ( ( 4 ( J − I ) − 2 B ) q 1 × q 1 Z q 1 × 1 ( i ) ) ⊗ J n × n J n × 1 + 2 I ~ q 1 n × q 1 n ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) ( 2 J p 2 × q 1 Z q 1 × 1 ( i ) ) ⊗ J 1 × n J n × 1 ) \begin{pmatrix}
O_{p_1\times1}\otimes nJ_{n\times1}\\
\left((4(J-I)-2B)_{q_1\times q_1}Z^{(i)}_{q_1\times1}\right)\otimes J_{n\times n}J_{n\times1}+2\widetilde{I}_{q_1n\times q_1n}\left(Z^{(i)}_{q_1\times1}\otimes J_{n\times1}\right)\\
\left(2J_{p_2\times q_1}Z^{(i)}_{q_1\times1}\right)\otimes J_{1\times n}J_{n\times1}
\end{pmatrix} O p 1 × 1 ⊗ n J n × 1 ( ( 4 ( J − I ) − 2 B ) q 1 × q 1 Z q 1 × 1 ( i ) ) ⊗ J n × n J n × 1 + 2 I q 1 n × q 1 n ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) ( 2 J p 2 × q 1 Z q 1 × 1 ( i ) ) ⊗ J 1 × n J n × 1
( O p 1 n × 1 O q 1 × 1 ⊗ n J n × 1 + ( Z q 1 × 1 ( i ) ⊗ 2 ( J − I ) n × n J n × 1 ) O p 2 × 1 ) \begin{pmatrix}
O_{p_1n\times1}\\
O_{q_1\times1}\otimes nJ_{n\times1}+\left(Z^{(i)}_{q_1\times1}\otimes2(J-I)_{n\times n}J_{n\times1}\right)\\
O_{p_2\times1}
\end{pmatrix} O p 1 n × 1 O q 1 × 1 ⊗ n J n × 1 + ( Z q 1 × 1 ( i ) ⊗ 2 ( J − I ) n × n J n × 1 ) O p 2 × 1
( O p 1 n × 1 O q 1 n × 1 + 2 ( Z q 1 × 1 ( i ) ⊗ ( n − 1 ) J n × 1 ) O p 2 × 1 ) \begin{pmatrix}
O_{p_1n\times1}\\
O_{q_1n\times1}+2\left(Z^{(i)}_{q_1\times1}\otimes(n-1)J_{n\times1}\right)\\
O_{p_2\times1}
\end{pmatrix} O p 1 n × 1 O q 1 n × 1 + 2 ( Z q 1 × 1 ( i ) ⊗ ( n − 1 ) J n × 1 ) O p 2 × 1
( O p 1 n × 1 O q 1 n × 1 + 2 ( n − 1 ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) O p 2 × 1 ) = 2 ( n − 1 ) χ i . \begin{pmatrix}
O_{p_1n\times1}\\
O_{q_1n\times1}+2(n-1)\left(Z^{(i)}_{q_1\times1}\otimes J_{n\times1}\right)\\
O_{p_2\times1}
\end{pmatrix}
=2(n-1)\chi_i. O p 1 n × 1 O q 1 n × 1 + 2 ( n − 1 ) ( Z q 1 × 1 ( i ) ⊗ J n × 1 ) O p 2 × 1 = 2 ( n − 1 ) χ i .
Thus, 2 ( n − 1 ) 2(n-1) 2 ( n − 1 ) is an eigenvalue of D ( S G 1 ( n , n ) ∨ ˙ G 2 ) D(SG_1(n,n)\mathbin{\dot\vee}G_2) D ( S G 1 ( n , n ) ∨ ˙ G 2 ) with multiplicity at least q 1 − p 1 q_1-p_1 q 1 − p 1 .
Now, let Y ( j ) Y^{(j)} Y ( j ) be an eigenvector of B ( G 1 ) B(G_1) B ( G 1 ) corresponding to the eigenvalue λ 1 j + r 2 − 2 \lambda_{1j+r_2-2} λ 1 j + r 2 − 2 , different from the regularity of L ( G 1 ) L(G_1) L ( G 1 ) . Then, Y ( j ) Y^{(j)} Y ( j ) is orthogonal to J J J and by Lemma 2.1 , J R ( G 1 ) = 2 J J R(G_1)=2J J R ( G 1 ) = 2 J .
Let ϕ j = ( ( R ( G 1 ) Y p 1 × q 1 ( j ) ⊗ J n × 1 ) p 1 n × 1 ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) q 1 n × 1 O p 2 × 1 ) \phi_j=\begin{pmatrix}(R(G_1)Y^{(j)}_{p_1\times q_1}\otimes J_{n\times1})_{p_1n\times1}\\(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})_{q_1n\times1}\\O_{p_2\times1}\end{pmatrix} ϕ j = ( R ( G 1 ) Y p 1 × q 1 ( j ) ⊗ J n × 1 ) p 1 n × 1 ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) q 1 n × 1 O p 2 × 1 , then
D ( S G 1 ( n , n ) ∨ ˙ G 2 ) ϕ j = ( ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) + ( 3 J − 2 R ) p 1 × q 1 ( − Y ( j ) ) q 1 × 1 ⊗ ( J n × n J n × 1 ) + 2 I ~ p 1 n × p 1 n ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) ( ( 3 J − 2 R T ) q 1 × p 1 ⊗ J n × n ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) + ( ( 4 J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) + 2 I ~ q 1 n × q 1 n ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) ( J p 2 × p 1 ⊗ J 1 × n ) ( R ( G 1 ) Y p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 ( J p 2 × q 1 ⊗ J 1 × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) ) = ( ( ( 2 ( J − I ) p 1 × p 1 R Y p 1 × 1 ( j ) ) + ( 3 J − 2 R ) p 1 × q 1 ( − Y ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( R Y p 1 × 1 ( j ) ⊗ 2 ( J − I ) n × n J n × 1 ) ( ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ( R ( G 1 ) Y ( j ) ) p 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( − Y ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) ( J p 2 × p 1 ( R ( G 1 ) Y ( j ) ) p 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) + 2 ( − J p 2 × q 1 Y q 1 × 1 ( j ) ) ⊗ ( J 1 × n J n × 1 ) ) = ( 2 ( n − 1 ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) 2 ( n − 1 ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) O p 2 × 1 ) = 2 ( n − 1 ) ϕ j . \begin{aligned}
D(SG_1(n,n)\mathbin{\dot\vee}G_2)\phi_j=
\begin{pmatrix}
(2(J-I)_{p_1\times p_1}\otimes J_{n\times n})(RY^{(j)}_{p_1\times1}\otimes J_{n\times1})+
(3J-2R)_{p_1\times q_1}(-Y^{(j)})_{q_1\times1}\otimes(J_{n\times n}J_{n\times1})+
2\widetilde{I}_{p_1n\times p_1n}(RY^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
((3J-2R^{T})_{q_1\times p_1}\otimes J_{n\times n})(RY^{(j)}_{p_1\times1}\otimes J_{n\times1})+
((4J-I)-2B(G_1))_{q_1\times q_1}\otimes J_{n\times n})(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})+
2\widetilde{I}_{q_1n\times q_1n}(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})\\
(J_{p_2\times p_1}\otimes J_{1\times n})(R(G_1)Y^{(j)}_{p_1\times1}\otimes J_{n\times1})+
2(J_{p_2\times q_1}\otimes J_{1\times n})(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})
\end{pmatrix} \\
&=
\begin{pmatrix}
\left((2(J-I)_{p_1\times p_1}RY^{(j)}_{p_1\times1})+(3J-2R)_{p_1\times q_1}(-Y^{(j)})_{q_1\times1}\right)\otimes(J_{n\times n}J_{n\times1})+
(RY^{(j)}_{p_1\times1}\otimes2(J-I)_{n\times n}J_{n\times1})\\
\left((3J-2R(G_1)^{T})_{q_1\times p_1}(R(G_1)Y^{(j)})_{p_1\times1}\right)\otimes(J_{n\times n}J_{n\times1})+
\left((4(J-I)-2B(G_1))_{q_1\times q_1}(-Y^{(j)})_{q_1\times1}\right)\otimes(J_{n\times n}J_{n\times1})+
\left(I_{q_1\times q_1}\otimes2(J-I)_{n\times n}\right)(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})\\
\left(J_{p_2\times p_1}(R(G_1)Y^{(j)})_{p_1\times1}\right)\otimes(J_{1\times n}J_{n\times1})+
2(-J_{p_2\times q_1}Y^{(j)}_{q_1\times1})\otimes(J_{1\times n}J_{n\times1})
\end{pmatrix} \\
&=
\begin{pmatrix}
2(n-1)(RY^{(j)}_{p_1\times1}\otimes J_{n\times1})\\
2(n-1)(-Y^{(j)}_{q_1\times1}\otimes J_{n\times1})\\
O_{p_2\times1}
\end{pmatrix} \\
&=2(n-1)\phi_j.
\end{aligned} D ( S G 1 ( n , n ) ∨ ˙ G 2 ) ϕ j = ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) + ( 3 J − 2 R ) p 1 × q 1 ( − Y ( j ) ) q 1 × 1 ⊗ ( J n × n J n × 1 ) + 2 I p 1 n × p 1 n ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) (( 3 J − 2 R T ) q 1 × p 1 ⊗ J n × n ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) + (( 4 J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) + 2 I q 1 n × q 1 n ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) ( J p 2 × p 1 ⊗ J 1 × n ) ( R ( G 1 ) Y p 1 × 1 ( j ) ⊗ J n × 1 ) + 2 ( J p 2 × q 1 ⊗ J 1 × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) = ( ( 2 ( J − I ) p 1 × p 1 R Y p 1 × 1 ( j ) ) + ( 3 J − 2 R ) p 1 × q 1 ( − Y ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( R Y p 1 × 1 ( j ) ⊗ 2 ( J − I ) n × n J n × 1 ) ( ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ( R ( G 1 ) Y ( j ) ) p 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ( − Y ( j ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) ( J p 2 × p 1 ( R ( G 1 ) Y ( j ) ) p 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) + 2 ( − J p 2 × q 1 Y q 1 × 1 ( j ) ) ⊗ ( J 1 × n J n × 1 ) = 2 ( n − 1 ) ( R Y p 1 × 1 ( j ) ⊗ J n × 1 ) 2 ( n − 1 ) ( − Y q 1 × 1 ( j ) ⊗ J n × 1 ) O p 2 × 1 = 2 ( n − 1 ) ϕ j .
Thus, 2 ( n − 1 ) 2(n-1) 2 ( n − 1 ) is an eigenvalue of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) with multiplicity at least ( q 1 − p 1 ) + ( p 1 − 1 ) = q 1 − 1 (q_1-p_1)+(p_1-1)=q_1-1 ( q 1 − p 1 ) + ( p 1 − 1 ) = q 1 − 1 .
Now, let λ 2 m ≠ r 2 \lambda_{2m}\ne r_2 λ 2 m = r 2 (for m = 2 , 3 , … , p 2 m=2,3,\ldots,p_2 m = 2 , 3 , … , p 2 ) be an eigenvalue of A ( G 2 ) A(G_2) A ( G 2 ) with an eigenvector W ( m ) W^{(m)} W ( m ) , and let η m = ( O p 1 n × 1 O q 1 n × 1 W ( m ) ) \eta_m=\begin{pmatrix} O_{p_1n\times1}\\ O_{q_1n\times1}\\ W^{(m)} \end{pmatrix} η m = O p 1 n × 1 O q 1 n × 1 W ( m ) . Then, D ( S G 1 ( n , n ) ∨ G 2 ) η m = − ( λ 2 m + 2 ) η m D(SG_1(n,n)\vee G_2)\eta_m=-(\lambda_{2m}+2)\eta_m D ( S G 1 ( n , n ) ∨ G 2 ) η m = − ( λ 2 m + 2 ) η m . Thus, η m \eta_m η m is an eigenvector of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) with an eigenvalue − ( λ 2 m + 2 ) -(\lambda_{2m}+2) − ( λ 2 m + 2 ) .
Thus, we have obtained ( n − 1 ) ( p 1 + q 1 ) + ( p 1 − 1 ) + ( q 1 − 1 ) + ( p 2 − 1 ) = n p 1 + n q 1 + p 2 − 3 (n-1)(p_1+q_1)+(p_1-1)+(q_1-1)+(p_2-1)=np_1+nq_1+p_2-3 ( n − 1 ) ( p 1 + q 1 ) + ( p 1 − 1 ) + ( q 1 − 1 ) + ( p 2 − 1 ) = n p 1 + n q 1 + p 2 − 3 eigenvalues. The remaining three eigenvalues are those of the equitable quotient matrix of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) ,
M = ( 2 ( n p 1 − 1 ) n ( 3 q 1 − 2 r 1 ) p 2 n ( 3 p 1 − 4 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 p 2 p 1 n 2 q 1 n 2 ( n − 1 ) − r 2 ) . M=\begin{pmatrix}
2(np_1-1) & n(3q_1-2r_1) & p_2\\
n(3p_1-4) & 4(q_1-r_1)n+2(n-1) & 2p_2\\
p_1n & 2q_1n & 2(n-1)-r_2
\end{pmatrix}. M = 2 ( n p 1 − 1 ) n ( 3 p 1 − 4 ) p 1 n n ( 3 q 1 − 2 r 1 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 q 1 n p 2 2 p 2 2 ( n − 1 ) − r 2 .
The following lemma establishes the distance between any two vertices of S G 1 ( n , n ) ∨ G 2 SG_1(n,n)\vee G_2 S G 1 ( n , n ) ∨ G 2 .
Proposition 3.3. Let V ( G ) = { v 1 , v 2 , … , v p } V(G)=\{v_1,v_2,\ldots,v_p\} V ( G ) = { v 1 , v 2 , … , v p } and E ( G ) = { e 1 , e 2 , … , e q } E(G)=\{e_1,e_2,\ldots,e_q\} E ( G ) = { e 1 , e 2 , … , e q } . Then, the distance between any two distinct vertices, say x x x and y y y , in S G 1 ( n , n ) ∨ G 2 SG_1(n,n)\vee G_2 S G 1 ( n , n ) ∨ G 2 is as follows,
d ( x , y ) = { 2 if x = ( v i , j ) , y = ( v l , s ) and v i is adjacent to v l in G 1 , 4 if x = ( v i , j ) , y = ( v l , s ) and v i is not adjacent to v l in G 1 , 1 if x = ( v i , j ) , y = ( e l , s ) and v i is incident with e l in G 1 , 3 if x = ( v i , j ) , y = ( e l , s ) and v i is not incident with e l in G 1 , 2 if x = ( v i , j ) and y ∈ V ( G 2 ) , 2 if x = ( e i , j ) , y = ( e l , s ) , 1 if x = ( e i , j ) and y ∈ V ( G 2 ) , 1 if x , y ∈ V ( G 2 ) and x , y are adjacent in V ( G 2 ) , 2 if x , y ∈ V ( G 2 ) and x , y are not adjacent in V ( G 2 ) . d(x,y)=\begin{cases}
2 & \text{if }x=(v_i,j),y=(v_l,s)\text{ and }v_i\text{ is adjacent to }v_l\text{ in }G_1,\\
4 & \text{if }x=(v_i,j),y=(v_l,s)\text{ and }v_i\text{ is not adjacent to }v_l\text{ in }G_1,\\
1 & \text{if }x=(v_i,j),y=(e_l,s)\text{ and }v_i\text{ is incident with }e_l\text{ in }G_1,\\
3 & \text{if }x=(v_i,j),y=(e_l,s)\text{ and }v_i\text{ is not incident with }e_l\text{ in }G_1,\\
2 & \text{if }x=(v_i,j)\text{ and }y\in V(G_2),\\
2 & \text{if }x=(e_i,j),y=(e_l,s),\\
1 & \text{if }x=(e_i,j)\text{ and }y\in V(G_2),\\
1 & \text{if }x,y\in V(G_2)\text{ and }x,y\text{ are adjacent in }V(G_2),\\
2 & \text{if }x,y\in V(G_2)\text{ and }x,y\text{ are not adjacent in }V(G_2).
\end{cases} d ( x , y ) = ⎩ ⎨ ⎧ 2 4 1 3 2 2 1 1 2 if x = ( v i , j ) , y = ( v l , s ) and v i is adjacent to v l in G 1 , if x = ( v i , j ) , y = ( v l , s ) and v i is not adjacent to v l in G 1 , if x = ( v i , j ) , y = ( e l , s ) and v i is incident with e l in G 1 , if x = ( v i , j ) , y = ( e l , s ) and v i is not incident with e l in G 1 , if x = ( v i , j ) and y ∈ V ( G 2 ) , if x = ( e i , j ) , y = ( e l , s ) , if x = ( e i , j ) and y ∈ V ( G 2 ) , if x , y ∈ V ( G 2 ) and x , y are adjacent in V ( G 2 ) , if x , y ∈ V ( G 2 ) and x , y are not adjacent in V ( G 2 ) .
Lemma 3.4. Let G j G_j G j be an r j r_j r j -regular graph on p j p_j p j vertices and q j q_j q j edges with the adjacency matrix A ( G j ) A(G_j) A ( G j ) and adjacency spectrum { r j , λ j 2 , λ j 3 , … , λ j p j } \{r_j,\lambda_{j2},\lambda_{j3},\ldots,\lambda_{jp_j}\} { r j , λ j 2 , λ j 3 , … , λ j p j } , j = 1 , 2 j=1,2 j = 1 , 2 . Then, the distance spectrum of S G 1 ( n , n ) ∨ G 2 SG_1(n,n)\vee G_2 S G 1 ( n , n ) ∨ G 2 consists of − 2 -2 − 2 with multiplicity ( n − 2 ) p 1 + n q 1 (n-2)p_1+nq_1 ( n − 2 ) p 1 + n q 1 ; − 2 − λ 2 m -2-\lambda_{2m} − 2 − λ 2 m , m = 2 , 3 , … , p 2 m=2,3,\ldots,p_2 m = 2 , 3 , … , p 2 ; − 2 ( n t + 1 ) -2(nt+1) − 2 ( n t + 1 ) , t = ( λ 1 i + 1 ) ± ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) 2 t=\frac{(\lambda_{1i}+1)\pm\sqrt{(\lambda_{1i}+1)^2+4(\lambda_{1i}+r_1)}}{2} t = 2 ( λ 1 i + 1 ) ± ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) , i = 2 , 3 , … , p 1 i=2,3,\ldots,p_1 i = 2 , 3 , … , p 1 ; and the eigenvalues of the matrix,
( ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) ( 3 q 1 − 2 r 1 ) n 2 p 2 n ( 3 p 1 − 4 ) 2 ( q 1 n − 1 ) p 2 2 p 1 n q 1 n 2 ( p 2 − 1 ) − r 2 ) . \begin{pmatrix}
(4(p_1-1)-2r_1)n+2(n-1) & (3q_1-2r_1)n & 2p_2\\
n(3p_1-4) & 2(q_1n-1) & p_2\\
2p_1n & q_1n & 2(p_2-1)-r_2
\end{pmatrix}. ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) n ( 3 p 1 − 4 ) 2 p 1 n ( 3 q 1 − 2 r 1 ) n 2 ( q 1 n − 1 ) q 1 n 2 p 2 p 2 2 ( p 2 − 1 ) − r 2 .
Proof. By suitable labeling of the vertices of ( S G 1 ( n , n ) ∨ G 2 ) (SG_1(n,n)\vee G_2) ( S G 1 ( n , n ) ∨ G 2 ) , the distance matrix D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) can be represented as
( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n + I ~ p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n 2 J p 1 n × p 2 ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) + I ~ q 1 n × q 1 n J q 1 n × p 2 2 J p 2 × p 1 n 2 J p 2 × q 1 n ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 ) . \begin{pmatrix}
(4(J-I)-2A(G_1))_{p_1\times p_1}\otimes J_{n\times n}+\widetilde I_{p_1n\times p_1n} & (3J-2R(G_1))_{p_1\times q_1}\otimes J_{n\times n} & 2J_{p_1n\times p_2}\\
(3J-2R(G_1)^{T})_{q_1\times p_1}\otimes J_{n\times n} & (2(J-I)_{q_1\times q_1}\otimes J_{n\times n})+\widetilde I_{q_1n\times q_1n} & J_{q_1n\times p_2}\\
2J_{p_2\times p_1n} & 2J_{p_2\times q_1n} & (2(J-I)-A(G_2))_{p_2\times p_2}
\end{pmatrix}. ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n + I p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n 2 J p 2 × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) + I q 1 n × q 1 n 2 J p 2 × q 1 n 2 J p 1 n × p 2 J q 1 n × p 2 ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 .
Let λ 1 i ≠ r 1 \lambda_{1i} \ne r_1 λ 1 i = r 1 be an eigenvalue of A ( G 1 ) A(G_1) A ( G 1 ) with an eigenvector X ( i ) X^{(i)} X ( i ) . Then, by Perron-Frobenius Theorem, X ( i ) X^{(i)} X ( i ) is orthogonal to J J J . Consider ϕ i = \phi_i= ϕ i =
( t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1 ) \begin{pmatrix}
tX^{(i)}_{p_1 \times1} \otimes J_{n \times1}\\
R(G_1)^{T}_{q_1 \times1}X^{(i)}_{p_1 \times1}\\
O_{p_2 \times1}
\end{pmatrix} t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1
is an eigenvector of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) . If μ i \mu_i μ i is an eigenvalue of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_1(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) with eigenvector ϕ i \phi_i ϕ i , then D ( S G 1 ( n , n ) ∨ G 2 ) ϕ i = μ i ϕ i D(SG_1(n,n)\vee G_2)\phi_i=\mu_i\phi_i D ( S G 1 ( n , n ) ∨ G 2 ) ϕ i = μ i ϕ i . That is,
( ( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n ) ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + I ~ p 1 n × p 1 n ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) ( ( 3 J − 2 R ( G 1 ) ) q 1 × p 1 ( t X ( i ) ) p 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) I ~ q 1 n × q 1 n ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) ( 2 J p 2 × p 1 ⊗ J 1 × n ) ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + ( J p 2 × q 1 ⊗ J 1 × 1 ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) ) = μ i ( t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1 ) \begin{aligned}
\begin{pmatrix}
\left((4(J-I)-2A(G_1))_{p_1\times p_1}\otimes J_{n\times n}\right)\left(tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\right)+
\widetilde I_{p_1n\times p_1n}\left(tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\right)+\\
\left((3J-2R(G_1))_{p_1\times q_1}\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\right)\otimes\left(J_{n\times n}J_{n\times1}\right)\\
\left((3J-2R(G_1))_{q_1\times p_1}\left(tX^{(i)}\right)_{p_1\times1}\right)\otimes\left(J_{n\times n}J_{n\times1}\right)+\\
\left(2(J-I)_{q_1\times q_1}\otimes J_{n\times n}\right)\left(\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\otimes J_{n\times1}\right)\\
\widetilde I_{q_1n\times q_1n}\left(\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\otimes J_{n\times1}\right)\\
\left(2J_{p_2\times p_1}\otimes J_{1\times n}\right)\left(tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\right)+\\
\left(J_{p_2\times q_1}\otimes J_{1\times1}\right)\left(\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\otimes J_{n\times1}\right)
\end{pmatrix}
=
\mu_i
\begin{pmatrix}
tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\\
R(G_1)^T_{q_1\times1}X^{(i)}_{p_1\times1}\\
O_{p_2\times1}
\end{pmatrix}
\end{aligned} ( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n ) ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + I p 1 n × p 1 n ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) ( ( 3 J − 2 R ( G 1 ) ) q 1 × p 1 ( t X ( i ) ) p 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) I q 1 n × q 1 n ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) ( 2 J p 2 × p 1 ⊗ J 1 × n ) ( t X p 1 × 1 ( i ) ⊗ J n × 1 ) + ( J p 2 × q 1 ⊗ J 1 × 1 ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ J n × 1 ) = μ i t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1
( ( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 t X p 1 × 1 ( i ) ) ⊗ ( J n × n J n × 1 ) + ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( I ~ p 1 × p 1 ⊗ 2 ( J − I ) n × n ) ( t X p 1 n × 1 ( i ) ⊗ J n × 1 ) ( − 2 t ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + ( 2 ( J − I ) R ( G 1 ) T ) q 1 × 1 ⊗ ( n J n × 1 ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ 2 ( J − I ) n × n J n × 1 ) 2 t ( ( J X ( i ) ) p 2 × 1 ⊗ n J 1 × 1 ) + ( J p 2 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) ) = μ i ( t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1 ) \begin{aligned}
\begin{pmatrix}
\left((4(J-I)-2A(G_1))_{p_1\times p_1}tX^{(i)}_{p_1\times1}\right)\otimes\left(J_{n\times n}J_{n\times1}\right)+\\
\left((3J-2R(G_1))_{p_1\times q_1}\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\right)\otimes\left(J_{n\times n}J_{n\times1}\right)+\\
\left(\widetilde I_{p_1\times p_1}\otimes2(J-I)_{n\times n}\right)\left(tX^{(i)}_{p_1n\times1}\otimes J_{n\times1}\right)\\
\left(-2t\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\right)\otimes\left(nJ_{n\times1}\right)+\left(2(J-I)R(G_1)^T\right)_{q_1\times1}\otimes\left(nJ_{n\times1}\right)\\
\left(\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\otimes2(J-I)_{n\times n}J_{n\times1}\right)\\
2t\left((JX^{(i)})_{p_2\times1}\otimes nJ_{1\times1}\right)+\\
\left(J_{p_2\times q_1}\left(R(G_1)^TX^{(i)}\right)_{q_1\times1}\right)\otimes\left(J_{1\times n}J_{n\times1}\right)
\end{pmatrix}
=
\mu_i
\begin{pmatrix}
tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\\
R(G_1)^T_{q_1\times1}X^{(i)}_{p_1\times1}\\
O_{p_2\times1}
\end{pmatrix}
\end{aligned} ( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 t X p 1 × 1 ( i ) ) ⊗ ( J n × n J n × 1 ) + ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J n × n J n × 1 ) + ( I p 1 × p 1 ⊗ 2 ( J − I ) n × n ) ( t X p 1 n × 1 ( i ) ⊗ J n × 1 ) ( − 2 t ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( n J n × 1 ) + ( 2 ( J − I ) R ( G 1 ) T ) q 1 × 1 ⊗ ( n J n × 1 ) ( ( R ( G 1 ) T X ( i ) ) q 1 × 1 ⊗ 2 ( J − I ) n × n J n × 1 ) 2 t ( ( J X ( i ) ) p 2 × 1 ⊗ n J 1 × 1 ) + ( J p 2 × q 1 ( R ( G 1 ) T X ( i ) ) q 1 × 1 ) ⊗ ( J 1 × n J n × 1 ) = μ i t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) q 1 × 1 T X p 1 × 1 ( i ) O p 2 × 1
( ( t ( − 2 n − 2 − 2 λ 1 i n ) − 2 n ( λ 1 i + r 1 ) ) ( X p 1 × 1 ( i ) ⊗ J n × 1 ) ( − 2 n t − 2 ) ( R ( G 1 ) T X p 1 × 1 ( i ) ) O p 2 × 1 ) = μ i ( t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) T X p 1 × 1 ( i ) O p 2 × 1 ) . \begin{pmatrix}
\left(t(-2n-2-2\lambda_{1i}n)-2n(\lambda_{1i}+r_1)\right)\left(X^{(i)}_{p_1\times1}\otimes J_{n\times1}\right)\\
(-2nt-2)\left(R(G_1)^TX^{(i)}_{p_1\times1}\right)\\
O_{p_2\times1}
\end{pmatrix}
=
\mu_i
\begin{pmatrix}
tX^{(i)}_{p_1\times1}\otimes J_{n\times1}\\
R(G_1)^TX^{(i)}_{p_1\times1}\\
O_{p_2\times1}
\end{pmatrix}. ( t ( − 2 n − 2 − 2 λ 1 i n ) − 2 n ( λ 1 i + r 1 ) ) ( X p 1 × 1 ( i ) ⊗ J n × 1 ) ( − 2 n t − 2 ) ( R ( G 1 ) T X p 1 × 1 ( i ) ) O p 2 × 1 = μ i t X p 1 × 1 ( i ) ⊗ J n × 1 R ( G 1 ) T X p 1 × 1 ( i ) O p 2 × 1 .
Thus,
t ( − 2 n − 2 − 2 n λ 1 i ) − 2 n ( λ 1 i + r 1 ) = μ i t − 2 n t − 2 = μ i . \begin{aligned}
t(-2n-2-2n\lambda_{1i})-2n(\lambda_{1i}+r_1)&=\mu_i t\\
-2nt-2&=\mu_i.
\end{aligned} t ( − 2 n − 2 − 2 n λ 1 i ) − 2 n ( λ 1 i + r 1 ) − 2 n t − 2 = μ i t = μ i .
Hence,
t 2 − t ( λ 1 i + 1 ) − ( λ 1 i + r 1 ) = 0. t^2-t(\lambda_{1i}+1)-(\lambda_{1i}+r_1)=0. t 2 − t ( λ 1 i + 1 ) − ( λ 1 i + r 1 ) = 0.
Therefore, t t t has two values, t 1 = ( λ 1 i + 1 ) + ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) 2 t_{1}=\frac{(\lambda_{1i}+1)+\sqrt{(\lambda_{1i}+1)^{2}+4(\lambda_{1i}+r_{1})}}{2} t 1 = 2 ( λ 1 i + 1 ) + ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) , t 2 = ( λ 1 i + 1 ) − ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) 2 t_{2}=\frac{(\lambda_{1i}+1)-\sqrt{(\lambda_{1i}+1)^{2}+4(\lambda_{1i}+r_{1})}}{2} t 2 = 2 ( λ 1 i + 1 ) − ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) .
That is, we get two values of μ i \mu_i μ i , − 2 ( n t 1 + 1 ) -2(nt_{1}+1) − 2 ( n t 1 + 1 ) and − 2 ( n t 2 + 1 ) -2(nt_{2}+1) − 2 ( n t 2 + 1 ) , corresponding to each eigenvalue λ 1 i ≠ r 1 \lambda_{1i}\ne r_{1} λ 1 i = r 1 of G 1 G_{1} G 1 . Thus, we obtain a total of 2 p 1 − 2 2p_{1}-2 2 p 1 − 2 eigenvalues of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_{1}(n,n)\vee G_{2}) D ( S G 1 ( n , n ) ∨ G 2 ) .
Now, consider the vector (1). Then we get,
D ( S G 1 ( n , n ) ∨ G 2 ) z ( s , t ) = − 2 z ( s , t ) , s = 1 , 2 , 3 , … , p 1 , t = 2 , 3 , … , n . D(SG_{1}(n,n)\vee G_{2})z^{(s,t)}=-2z^{(s,t)},\ s=1,2,3,\ldots,p_{1},\ t=2,3,\ldots,n. D ( S G 1 ( n , n ) ∨ G 2 ) z ( s , t ) = − 2 z ( s , t ) , s = 1 , 2 , 3 , … , p 1 , t = 2 , 3 , … , n .
This yields ( n − 1 ) p 1 (n-1)p_{1} ( n − 1 ) p 1 linearly independent eigenvectors of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_{1}(n,n)\vee G_{2}) D ( S G 1 ( n , n ) ∨ G 2 ) associated with the eigenvalue − 2 -2 − 2 . Similarly, for the vector (2), we have
D ( S G 1 ( n , n ) ∨ G 2 ) h ( k , l ) = − 2 h ( k , l ) , k = 1 , 2 , … , q 1 , l = 2 , 3 , … , n . D(SG_{1}(n,n)\vee G_{2})h^{(k,l)}=-2h^{(k,l)},\ k=1,2,\ldots,q_{1},l=2,3,\ldots,n. D ( S G 1 ( n , n ) ∨ G 2 ) h ( k , l ) = − 2 h ( k , l ) , k = 1 , 2 , … , q 1 , l = 2 , 3 , … , n .
Thereby, we obtain a further collection of q 1 ( n − 1 ) q_{1}(n-1) q 1 ( n − 1 ) linearly independent eigenvectors corresponding to the eigenvalue − 2 -2 − 2 .
Let Z q 1 × 1 ( d ) Z^{(d)}_{q_{1}\times1} Z q 1 × 1 ( d ) be an eigenvector of B ( G 1 ) B(G_{1}) B ( G 1 ) corresponding to the eigenvalue − 2 -2 − 2 , with multiplicity q 1 − p 1 q_{1}-p_{1} q 1 − p 1 . By Lemma 2.2, R ( G 1 ) p 1 × q 1 Z q 1 × 1 ( d ) = O R(G_{1})_{p_{1}\times q_{1}}Z^{(d)}_{q_{1}\times1}=O R ( G 1 ) p 1 × q 1 Z q 1 × 1 ( d ) = O . Let ψ d = ( O p 1 × 1 Z q 1 × 1 ( d ) ⊗ J n × 1 O p 2 × 1 ) \psi_{d}=\begin{pmatrix}O_{p_{1}\times1}\\ Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1}\\ O_{p_{2}\times1}\end{pmatrix} ψ d = O p 1 × 1 Z q 1 × 1 ( d ) ⊗ J n × 1 O p 2 × 1 . Then,
D ( S G 1 ( n , n ) ∨ G 2 ) ψ d = ( ( ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J n × n J n × 1 ) ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) + 2 I ~ q 1 n × q 1 n ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) ( J p 2 × q 1 ⊗ J 1 × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) ) = ( O p 1 × 1 ⊗ n J n × 1 ( 2 ( J − I ) q 1 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J n × n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) ( J p 2 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J 1 × n J n × 1 ) ) = ( O p 1 × 1 ⊗ J n × 1 ( − 2 n Z q 1 × 1 ( d ) ⊗ J n × 1 ) + ( Z q 1 × 1 ( d ) ⊗ ( 2 ( J − I ) n × n J n × 1 ) ) O p 2 × 1 ⊗ J 1 × 1 ) = ( O p 1 × 1 ⊗ J n × 1 − 2 ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) O p 2 × 1 ) = − 2 ψ d , \begin{aligned}
D(SG_{1}(n,n)\vee G_{2})\psi_{d}
&=\begin{pmatrix}
((3J-2R(G_{1}))_{p_{1}\times q_{1}}Z^{(d)}_{q_{1}\times1})\otimes(J_{n\times n}J_{n\times1})\\
(2(J-I)_{q_{1}\times q_{1}}\otimes J_{n\times n})(Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1})+2\widetilde{I}_{q_{1}n\times q_{1}n}(Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1})\\
(J_{p_{2}\times q_{1}}\otimes J_{1\times n})(Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1})
\end{pmatrix}\\
&=\begin{pmatrix}
O_{p_{1}\times1}\otimes nJ_{n\times1}\\
(2(J-I)_{q_{1}\times q_{1}}Z^{(d)}_{q_{1}\times1})\otimes(J_{n\times n}J_{n\times1})+(I_{q_{1}\times q_{1}}\otimes2(J-I)_{n\times n})(Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1})\\
(J_{p_{2}\times q_{1}}Z^{(d)}_{q_{1}\times1})\otimes(J_{1\times n}J_{n\times1})
\end{pmatrix}\\
&=\begin{pmatrix}
O_{p_{1}\times1}\otimes J_{n\times1}\\
(-2nZ^{(d)}_{q_{1}\times1}\otimes J_{n\times1})+(Z^{(d)}_{q_{1}\times1}\otimes(2(J-I)_{n\times n}J_{n\times1}))\\
O_{p_{2}\times1}\otimes J_{1\times1}
\end{pmatrix}\\
&=\begin{pmatrix}
O_{p_{1}\times1}\otimes J_{n\times1}\\
-2(Z^{(d)}_{q_{1}\times1}\otimes J_{n\times1})\\
O_{p_{2}\times1}
\end{pmatrix}
=-2\psi_{d},
\end{aligned} D ( S G 1 ( n , n ) ∨ G 2 ) ψ d = (( 3 J − 2 R ( G 1 ) ) p 1 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J n × n J n × 1 ) ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) + 2 I q 1 n × q 1 n ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) ( J p 2 × q 1 ⊗ J 1 × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) = O p 1 × 1 ⊗ n J n × 1 ( 2 ( J − I ) q 1 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J n × n J n × 1 ) + ( I q 1 × q 1 ⊗ 2 ( J − I ) n × n ) ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) ( J p 2 × q 1 Z q 1 × 1 ( d ) ) ⊗ ( J 1 × n J n × 1 ) = O p 1 × 1 ⊗ J n × 1 ( − 2 n Z q 1 × 1 ( d ) ⊗ J n × 1 ) + ( Z q 1 × 1 ( d ) ⊗ ( 2 ( J − I ) n × n J n × 1 )) O p 2 × 1 ⊗ J 1 × 1 = O p 1 × 1 ⊗ J n × 1 − 2 ( Z q 1 × 1 ( d ) ⊗ J n × 1 ) O p 2 × 1 = − 2 ψ d ,
which again yields q 1 − p 1 q_{1}-p_{1} q 1 − p 1 linearly independent eigenvectors corresponding to the eigenvalue − 2 -2 − 2 .
Finally, the pairwise disjointness of the coordinate support of the vectors { z ( s , t ) } \{z^{(s,t)}\} { z ( s , t ) } , { h ( k , l ) } \{h^{(k,l)}\} { h ( k , l ) } and { ψ d } \{\psi_{d}\} { ψ d } , (for all $s,t,k,l,d), ensures that they form a linearly independent set. Hence, − 2 -2 − 2 is an eigenvalue of D ( S G 1 ( n , n ) ∨ G 2 ) D(SG_{1}(n,n)\vee G_{2}) D ( S G 1 ( n , n ) ∨ G 2 ) with multiplicity at least ( n − 1 ) p 1 + ( n − 1 ) q 1 + ( q 1 − p 1 ) = ( n − 2 ) p 1 + n q 1 (n-1)p_{1}+(n-1)q_{1}+(q_{1}-p_{1})=(n-2)p_{1}+nq_{1} ( n − 1 ) p 1 + ( n − 1 ) q 1 + ( q 1 − p 1 ) = ( n − 2 ) p 1 + n q 1 .
Now, let λ 2 m ≠ r 2 \lambda_{2m}\ne r_{2} λ 2 m = r 2 (for m = 2 , 3 , … , p 2 m=2,3,\ldots,p_{2} m = 2 , 3 , … , p 2 ) be an eigenvalue of A ( G 2 ) A(G_{2}) A ( G 2 ) with an eigenvector W p 2 × 1 ( m ) W^{(m)}_{p_{2}\times1} W p 2 × 1 ( m ) .
Let χ m = ( O p 1 n × 1 O q 1 n × 1 W p 2 × 1 ( m ) ) \chi_{m}=\begin{pmatrix}O_{p_{1}n\times1}\\ O_{q_{1}n\times1}\\ W^{(m)}_{p_{2}\times1}\end{pmatrix} χ m = O p 1 n × 1 O q 1 n × 1 W p 2 × 1 ( m ) . Then,
D ( S G 1 ( n , n ) ∨ G 2 ) χ m = ( 2 J p 1 n × p 2 W p 2 × 1 ( m ) J q 1 n × p 2 W p 2 × 1 ( m ) ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 W p 2 × 1 ( m ) ) D(SG_{1}(n,n)\vee G_{2})\chi_{m}=
\begin{pmatrix}
2J_{p_{1}n\times p_{2}}W^{(m)}_{p_{2}\times1}\\
J_{q_{1}n\times p_{2}}W^{(m)}_{p_{2}\times1}\\
(2(J-I)-A(G_{2}))_{p_{2}\times p_{2}}W^{(m)}_{p_{2}\times1}
\end{pmatrix} D ( S G 1 ( n , n ) ∨ G 2 ) χ m = 2 J p 1 n × p 2 W p 2 × 1 ( m ) J q 1 n × p 2 W p 2 × 1 ( m ) ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 W p 2 × 1 ( m )
( O p 1 n × 1 O q 1 n × 1 ( − 2 − λ 2 m ) W p 2 × 1 ( m ) ) = ( − 2 − λ 2 m ) χ m . \begin{pmatrix}
O_{p_1n\times1}\\
O_{q_1n\times1}\\
(-2-\lambda_{2m})W^{(m)}_{p_2\times1}
\end{pmatrix}
=(-2-\lambda_{2m})\chi_m. O p 1 n × 1 O q 1 n × 1 ( − 2 − λ 2 m ) W p 2 × 1 ( m ) = ( − 2 − λ 2 m ) χ m .
This shows that ( − 2 − λ 2 m ) (-2-\lambda_{2m}) ( − 2 − λ 2 m ) is an eigenvalue of D ( S G 1 ( n , n ) ∨ ˙ G 2 ) D(S_{G_1}(n,n)\dot{\vee}G_2) D ( S G 1 ( n , n ) ∨ ˙ G 2 ) corresponding to the eigenvalue λ 2 m \lambda_{2m} λ 2 m of A ( G 2 ) A(G_2) A ( G 2 ) .
Thus, we get p 1 n + q 1 n + p 2 − 3 p_1n+q_1n+p_2-3 p 1 n + q 1 n + p 2 − 3 eigenvalues. The remaining 3 3 3 eigenvalues are given by the eigenvalues of the equitable quotient matrix of D ( S G 1 ( n , n ) ∨ G 2 ) D(S_{G_1}(n,n)\vee G_2) D ( S G 1 ( n , n ) ∨ G 2 ) ,
M = ( ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) ( 3 q 1 − 2 r 1 ) n 2 p 2 n ( 3 p 1 − 4 ) 2 ( q 1 n − 1 ) p 2 2 p 1 n q 1 n 2 ( p 2 − 1 ) − r 2 ) . M=\begin{pmatrix}
(4(p_1-1)-2r_1)n+2(n-1) & (3q_1-2r_1)n & 2p_2\\
n(3p_1-4) & 2(q_1n-1) & p_2\\
2p_1n & q_1n & 2(p_2-1)-r_2
\end{pmatrix}. M = ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) n ( 3 p 1 − 4 ) 2 p 1 n ( 3 q 1 − 2 r 1 ) n 2 ( q 1 n − 1 ) q 1 n 2 p 2 p 2 2 ( p 2 − 1 ) − r 2 .
Applications
In this section, we construct some new families of D D D -equienergetic graphs and D D D -cospectral graphs of diameter 4 4 4 .
Lemmas 3.2 and 3.4 demonstrate that the D D D -spectrum of S G 1 ( n , n ) ∨ ˙ G 2 S_{G_1}(n,n)\dot{\vee}G_2 S G 1 ( n , n ) ∨ ˙ G 2 and S G 1 ( n , n ) ∨ G 2 S_{G_1}(n,n)\vee G_2 S G 1 ( n , n ) ∨ G 2 , are determined by the orders, regularity and the adjacency spectra of G 1 G_1 G 1 and G 2 G_2 G 2 , together with the eigenvalue − 2 -2 − 2 of B ( G 1 ) B(G_1) B ( G 1 ) , the adjacency matrix of the line graph of G 1 G_1 G 1 . In view of this, Lemmas 3.2 and 3.4 yield the following theorems.
Theorem 4.1. Let G 1 G_1 G 1 and G 2 G_2 G 2 be regular cospectral graphs of the same order and regularity, and H H H be any regular graph. Then,
(i) S G 1 ( n , n ) ∨ ˙ H S_{G_1}(n,n)\dot{\vee}H S G 1 ( n , n ) ∨ ˙ H and S G 2 ( n , n ) ∨ ˙ H S_{G_2}(n,n)\dot{\vee}H S G 2 ( n , n ) ∨ ˙ H are D D D -cospectral.
(ii) S G 1 ( n , n ) ∨ H S_{G_1}(n,n)\vee H S G 1 ( n , n ) ∨ H and S G 2 ( n , n ) ∨ H S_{G_2}(n,n)\vee H S G 2 ( n , n ) ∨ H are D D D -cospectral.
Theorem 4.2. Let G G G be an r r r -regular graph on p p p vertices and q q q edges. For a fixed t ∈ N t\in\mathbb{N} t ∈ N , let T \mathcal{T} T be the family of all integer partitions T = ( t 1 , t 2 , … , t s ) \mathcal{T}=(t_1,t_2,\ldots,t_s) T = ( t 1 , t 2 , … , t s ) of t t t into parts t i ≥ 3 t_i\ge3 t i ≥ 3 (1 ≤ i ≤ s 1\le i\le s 1 ≤ i ≤ s ), and for each such T ∈ T \mathcal{T}\in\mathcal{T} T ∈ T , let
C T = C t 1 ∪ C t 2 ∪ ⋯ ∪ C t s C_{\mathcal{T}}=C_{t_1}\cup C_{t_2}\cup\cdots\cup C_{t_s} C T = C t 1 ∪ C t 2 ∪ ⋯ ∪ C t s be the corresponding disjoint union of cycles. Then,
(i) the graphs S G ( n , n ) ∨ ˙ C T S_G(n,n)\dot{\vee}C_{\mathcal{T}} S G ( n , n ) ∨ ˙ C T , T ∈ T \mathcal{T}\in\mathcal{T} T ∈ T , form a family of D D D -equienergetic graphs,
(ii) the graphs S G ( n , n ) ∨ C T S_G(n,n)\vee C_{\mathcal{T}} S G ( n , n ) ∨ C T , T ∈ T \mathcal{T}\in\mathcal{T} T ∈ T , form a family of D D D -equienergetic graphs.
Proof. Let λ 1 = r 1 ≥ λ 2 ≥ … ≥ λ p \lambda_1=r_1\ge\lambda_2\ge\ldots\ge\lambda_p λ 1 = r 1 ≥ λ 2 ≥ … ≥ λ p and μ 1 = 2 ≥ μ 2 ≥ … ≥ μ t \mu_1=2\ge\mu_2\ge\ldots\ge\mu_t μ 1 = 2 ≥ μ 2 ≥ … ≥ μ t be the adjacency eigenvalues of G G G and C T C_{\mathcal{T}} C T respectively.
(i) Then, by Lemma 3.2, the D D D -eigenvalues of S G ( n , n ) ∨ ˙ C T S_G(n,n)\dot{\vee}C_{\mathcal{T}} S G ( n , n ) ∨ ˙ C T are, − 2 -2 − 2 with multiplicity ( n − 1 ) ( p + q ) (n-1)(p+q) ( n − 1 ) ( p + q ) ; 2 ( n − 1 ) 2(n-1) 2 ( n − 1 ) with multiplicity q − 1 q-1 q − 1 ; − 2 ( n λ j + n r + 1 ) -2(n\lambda_j+nr+1) − 2 ( n λ j + n r + 1 ) , j = 2 , 3 , … , p j=2,3,\ldots,p j = 2 , 3 , … , p ; − 2 − μ i -2-\mu_i − 2 − μ i , i = 2 , 3 , … , t i=2,3,\ldots,t i = 2 , 3 , … , t ; the eigenvalues of the matrix,
( 2 ( n p − 1 ) n ( 3 q − 2 r ) t n ( 3 p − 4 ) 4 ( q − r ) n + 2 ( n − 1 ) 2 t p n 2 q n 2 ( t − 1 ) − 2 ) , \begin{pmatrix}
2(np-1) & n(3q-2r) & t\\
n(3p-4) & 4(q-r)n+2(n-1) & 2t\\
pn & 2qn & 2(t-1)-2
\end{pmatrix}, 2 ( n p − 1 ) n ( 3 p − 4 ) p n n ( 3 q − 2 r ) 4 ( q − r ) n + 2 ( n − 1 ) 2 q n t 2 t 2 ( t − 1 ) − 2 ,
for every partition T T T of t t t . Since μ i ≥ − 2 \mu_i \ge-2 μ i ≥ − 2 , for i = 2 , 3 , … , t i=2,3,\ldots,t i = 2 , 3 , … , t , we have, − ( μ i + 2 ) ≤ 0 -(\mu_i+2) \le0 − ( μ i + 2 ) ≤ 0 . Therefore,
∑ i = 2 t ∣ − 2 − μ i ∣ = ∑ i = 2 t ( 2 + μ i ) = 2 ( t − 1 ) + ∑ i = 2 t μ i = 2 t − 4. \begin{aligned}
\sum_{i=2}^{t} \left|-2-\mu_i\right|=\sum_{i=2}^{t}(2+\mu_i) \\
&=2(t-1)+\sum_{i=2}^{t}\mu_i \\
&=2t-4.
\end{aligned} i = 2 ∑ t ∣ − 2 − μ i ∣ = i = 2 ∑ t ( 2 + μ i ) = 2 ( t − 1 ) + i = 2 ∑ t μ i = 2 t − 4.
Thus, for every partitions T i , T j ∈ T T_i,T_j \in\mathcal{T} T i , T j ∈ T ,
E D ( S G ( n , n ) ∨ C T i ) = E D ( S G ( n , n ) ∨ C T j ) . E_D(SG(n,n)\vee CT_i)=E_D(SG(n,n)\vee CT_j). E D ( S G ( n , n ) ∨ C T i ) = E D ( S G ( n , n ) ∨ C T j ) .
(ii) By Lemma 3.2, the D D D -eigenvalues of S G ( n , n ) ∨ C T SG(n,n)\vee CT S G ( n , n ) ∨ C T are, − 2 -2 − 2 with multiplicity ( n − 2 ) p + n q (n-2)p+nq ( n − 2 ) p + n q ; − 2 − μ i -2-\mu_i − 2 − μ i , i = 2 , 3 , … , t i=2,3,\ldots,t i = 2 , 3 , … , t ; − 2 ( n t j + 1 ) -2(nt_j+1) − 2 ( n t j + 1 ) , t = ( λ i + 1 ) ± ( λ i + 1 ) 2 + 4 ( λ i + r 1 ) 2 t=\frac{(\lambda_i+1)\pm\sqrt{(\lambda_i+1)^2+4(\lambda_i+r_1)}}{2} t = 2 ( λ i + 1 ) ± ( λ i + 1 ) 2 + 4 ( λ i + r 1 ) , j = 1 , 2 j=1,2 j = 1 , 2 ; and the eigenvalues of the matrix
( ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) ( 3 q 1 − 2 r 1 ) n 2 t n ( 3 p 1 − 4 ) 2 ( q 1 n − 1 ) t 2 p 1 n q 1 n 2 ( t − 1 ) − 2 ) \begin{pmatrix}
(4(p_1-1)-2r_1)n+2(n-1) & (3q_1-2r_1)n & 2t \\
n(3p_1-4) & 2(q_1n-1) & t \\
2p_1n & q_1n & 2(t-1)-2
\end{pmatrix} ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) n ( 3 p 1 − 4 ) 2 p 1 n ( 3 q 1 − 2 r 1 ) n 2 ( q 1 n − 1 ) q 1 n 2 t t 2 ( t − 1 ) − 2
for every partition T T T of t t t . Therefore, as in part ( i ) (i) ( i ) ) we get,
∑ i = 2 t ∣ − 2 − μ i ∣ = 2 t − 4. \sum_{i=2}^{t}\left|-2-\mu_i\right|=2t-4. i = 2 ∑ t ∣ − 2 − μ i ∣ = 2 t − 4.
Hence,
E D ( S G ( n , n ) ∨ C T i ) = E D ( S G ( n , n ) ∨ C T j ) , E_D(SG(n,n)\vee CT_i)=E_D(SG(n,n)\vee CT_j), E D ( S G ( n , n ) ∨ C T i ) = E D ( S G ( n , n ) ∨ C T j ) ,
for every partitions T i , T j ∈ T T_i,T_j \in\mathcal{T} T i , T j ∈ T .
Table 1 lists some families of D D D -equienergetic graphs constructed using the generalized subdivision vertex join and generalized subdivision edge join of graphs.
Table 1. Some D D D -equienergetic graphs of diameter 4.
Lemmas 3.2 and 3.4 determine the distance spectra of S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\dot\vee(G_2\cup G_3) S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) and S G 1 ( n , n ) ∨ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\vee(G_2\cup G_3) S G 1 ( n , n ) ∨ ( G 2 ∪ G 3 ) , respectively, when G 2 G_2 G 2 and G 3 G_3 G 3 are regular graphs of the same degree.
The following theorems extend these results to the more general case in which G 2 G_2 G 2 and G 3 G_3 G 3 are regular graphs of distinct degrees.
Theorem 4.3. For i = 1 , 2 , 3 i=1,2,3 i = 1 , 2 , 3 , let G i G_i G i be an r i r_i r i -regular graph with p i p_i p i vertices and q i q_i q i edges whose adjacency spectrum is given by { r i , λ i 1 , λ i 2 , … , λ i p i } \{r_i,\lambda_{i1},\lambda_{i2},\ldots,\lambda_{ip_i}\} { r i , λ i 1 , λ i 2 , … , λ i p i } . If r 2 ≠ r 3 r_2\ne r_3 r 2 = r 3 , then the D D D -spectrum of S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\dot\vee(G_2\cup G_3) S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) consists of − 2 -2 − 2 with multiplicity ( n − 1 ) ( p 1 + q 1 ) (n-1)(p_1+q_1) ( n − 1 ) ( p 1 + q 1 ) ; 2 ( n − 1 ) 2(n-1) 2 ( n − 1 ) with multiplicity q 1 − 1 q_1-1 q 1 − 1 ; − 2 − λ i j -2-\lambda_{ij} − 2 − λ ij , i = 1 , 2 i=1,2 i = 1 , 2 ; j = 2 , 3 , … , p 2 j=2,3,\ldots,p_2 j = 2 , 3 , … , p 2 ; and the eigenvalues of the matrix,
( 2 ( n p 1 − 1 ) n ( 3 q 1 − 2 r 1 ) p 2 p 3 n ( 3 p 1 − 4 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 p 2 2 p 3 p 1 n 2 q 1 n 2 ( p 2 − 1 ) − r 2 2 p 3 p 1 n 2 q 1 n 2 p 2 2 ( p 3 − 1 ) − r 3 ) . \begin{pmatrix}
2(np_1-1) & n(3q_1-2r_1) & p_2 & p_3 \\
n(3p_1-4) & 4(q_1-r_1)n+2(n-1) & 2p_2 & 2p_3 \\
p_1n & 2q_1n & 2(p_2-1)-r_2 & 2p_3 \\
p_1n & 2q_1n & 2p_2 & 2(p_3-1)-r_3
\end{pmatrix}. 2 ( n p 1 − 1 ) n ( 3 p 1 − 4 ) p 1 n p 1 n n ( 3 q 1 − 2 r 1 ) 4 ( q 1 − r 1 ) n + 2 ( n − 1 ) 2 q 1 n 2 q 1 n p 2 2 p 2 2 ( p 2 − 1 ) − r 2 2 p 2 p 3 2 p 3 2 p 3 2 ( p 3 − 1 ) − r 3 .
Proof. By an appropriate labeling of vertices of S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\dot\vee(G_2\cup G_3) S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) , the distance matrix of D ( S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) ) D(S_{G_1}(n,n)\dot\vee(G_2\cup G_3)) D ( S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 )) can be represented as
( ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) + 2 I ~ p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n J p 1 n × p 2 J p 1 n × p 3 ( ( 3 J − 2 R ( G 1 ) ) q 1 × p 1 ⊗ J n × n ) ( ( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) + 2 I ~ q 1 n × q 1 n 2 J q 1 n × p 2 2 J q 1 n × p 3 J p 2 × p 1 n 2 J p 2 × q 1 n ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 J p 2 × p 3 J p 3 × p 1 n 2 J p 3 × q 1 n 2 J p 3 × p 2 ( 2 ( J − I ) − A ( G 3 ) ) p 3 × p 3 ) . \begin{pmatrix}
(2(J-I)_{p_1\times p_1}\otimes J_{n\times n})+2\widetilde I_{p_1n\times p_1n} & (3J-2R(G_1))_{p_1\times q_1}\otimes J_{n\times n} & J_{p_1n\times p_2} & J_{p_1n\times p_3} \\
((3J-2R(G_1))_{q_1\times p_1}\otimes J_{n\times n}) & ((4(J-I)-2B(G_1))_{q_1\times q_1}\otimes J_{n\times n})+2\widetilde I_{q_1n\times q_1n} & 2J_{q_1n\times p_2} & 2J_{q_1n\times p_3} \\
J_{p_2\times p_1n} & 2J_{p_2\times q_1n} & (2(J-I)-A(G_2))_{p_2\times p_2} & J_{p_2\times p_3} \\
J_{p_3\times p_1n} & 2J_{p_3\times q_1n} & 2J_{p_3\times p_2} & (2(J-I)-A(G_3))_{p_3\times p_3}
\end{pmatrix}. ( 2 ( J − I ) p 1 × p 1 ⊗ J n × n ) + 2 I p 1 n × p 1 n (( 3 J − 2 R ( G 1 ) ) q 1 × p 1 ⊗ J n × n ) J p 2 × p 1 n J p 3 × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n (( 4 ( J − I ) − 2 B ( G 1 ) ) q 1 × q 1 ⊗ J n × n ) + 2 I q 1 n × q 1 n 2 J p 2 × q 1 n 2 J p 3 × q 1 n J p 1 n × p 2 2 J q 1 n × p 2 ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 2 J p 3 × p 2 J p 1 n × p 3 2 J q 1 n × p 3 J p 2 × p 3 ( 2 ( J − I ) − A ( G 3 ) ) p 3 × p 3 .
The proof is analogous to that of Lemma 3.2.
Theorem 4.4. For i = 1 , 2 , 3 i=1,2,3 i = 1 , 2 , 3 , let G i G_i G i be an r i r_i r i -regular graph with order p i p_i p i and size q i q_i q i whose adjacency spectrum is given by { r i , λ i 1 , λ i 2 , … , λ i p i } \{r_i,\lambda_{i1},\lambda_{i2},\ldots,\lambda_{ip_i}\} { r i , λ i 1 , λ i 2 , … , λ i p i } . If r 2 ≠ r 3 r_2\ne r_3 r 2 = r 3 , then the D D D -spectrum of S G 1 ( n , n ) ∨ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\vee(G_2\cup G_3) S G 1 ( n , n ) ∨ ( G 2 ∪ G 3 ) consists of − 2 -2 − 2 with multiplicity ( n − 2 ) p 1 + n q 1 (n-2)p_1+nq_1 ( n − 2 ) p 1 + n q 1 ; − 2 − λ i j -2-\lambda_{ij} − 2 − λ ij , j = 2 , 3 , … , p 2 j=2,3,\ldots,p_2 j = 2 , 3 , … , p 2 , i = 2 , 3 i=2,3 i = 2 , 3 ; − 2 ( n t + 1 ) -2(nt+1) − 2 ( n t + 1 ) , t = ( λ 1 i + 1 ) ± ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) 2 t=\frac{(\lambda_{1i}+1)\pm\sqrt{(\lambda_{1i}+1)^2+4(\lambda_{1i}+r_1)}}{2} t = 2 ( λ 1 i + 1 ) ± ( λ 1 i + 1 ) 2 + 4 ( λ 1 i + r 1 ) , i = 2 , 3 , … , p 1 i=2,3,\ldots,p_1 i = 2 , 3 , … , p 1 ; and the eigenvalues of the matrix,
( ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) ( 3 q 1 − 2 r 1 ) n 2 p 2 2 p 3 n ( 3 p 1 − 4 ) 2 ( q 1 n − 1 ) p 2 p 3 2 p 1 n q 1 n 2 ( p 2 − 1 ) − r 2 2 p 3 2 p 1 n q 1 n 2 p 2 2 ( p 3 − 1 ) − r 3 ) . \begin{pmatrix}
(4(p_1-1)-2r_1)n+2(n-1) & (3q_1-2r_1)n & 2p_2 & 2p_3 \\
n(3p_1-4) & 2(q_1n-1) & p_2 & p_3 \\
2p_1n & q_1n & 2(p_2-1)-r_2 & 2p_3 \\
2p_1n & q_1n & 2p_2 & 2(p_3-1)-r_3
\end{pmatrix}. ( 4 ( p 1 − 1 ) − 2 r 1 ) n + 2 ( n − 1 ) n ( 3 p 1 − 4 ) 2 p 1 n 2 p 1 n ( 3 q 1 − 2 r 1 ) n 2 ( q 1 n − 1 ) q 1 n q 1 n 2 p 2 p 2 2 ( p 2 − 1 ) − r 2 2 p 2 2 p 3 p 3 2 p 3 2 ( p 3 − 1 ) − r 3 .
Proof. By an appropriate labeling of vertices of S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) S_{G_1}(n,n)\dot{\vee}(G_2\cup G_3) S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) , the distance matrix of D ( S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 ) ) D(S_{G_1}(n,n)\dot{\vee}(G_2\cup G_3)) D ( S G 1 ( n , n ) ∨ ˙ ( G 2 ∪ G 3 )) can be represented as
( ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n + I ~ p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n 2 J p 1 n × p 2 2 J p 1 n × p 3 ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) + I ~ q 1 n × q 1 n J q 1 n × p 2 J q 1 n × p 3 2 J p 2 × p 1 n 2 J p 2 × q 1 n ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 2 J p 2 × p 3 2 J p 3 × p 1 n J p 3 × q 1 n 2 J p 3 × p 2 ( 2 ( J − I ) − A ( G 3 ) ) p 3 × p 3 ) \begin{pmatrix}
\left(4(J-I)-2A(G_1)\right)_{p_1\times p_1}\otimes J_{n\times n}+\widetilde{I}_{p_1n\times p_1n} & \left(3J-2R(G_1)\right)_{p_1\times q_1}\otimes J_{n\times n} & 2J_{p_1n\times p_2} & 2J_{p_1n\times p_3} \\
\left(3J-2R(G_1)^{T}\right)_{q_1\times p_1}\otimes J_{n\times n} & \left(2(J-I)_{q_1\times q_1}\otimes J_{n\times n}\right)+\widetilde{I}_{q_1n\times q_1n} & J_{q_1n\times p_2} & J_{q_1n\times p_3} \\
2J_{p_2\times p_1n} & 2J_{p_2\times q_1n} & \left(2(J-I)-A(G_2)\right)_{p_2\times p_2} & 2J_{p_2\times p_3} \\
2J_{p_3\times p_1n} & J_{p_3\times q_1n} & 2J_{p_3\times p_2} & \left(2(J-I)-A(G_3)\right)_{p_3\times p_3}
\end{pmatrix} ( 4 ( J − I ) − 2 A ( G 1 ) ) p 1 × p 1 ⊗ J n × n + I p 1 n × p 1 n ( 3 J − 2 R ( G 1 ) T ) q 1 × p 1 ⊗ J n × n 2 J p 2 × p 1 n 2 J p 3 × p 1 n ( 3 J − 2 R ( G 1 ) ) p 1 × q 1 ⊗ J n × n ( 2 ( J − I ) q 1 × q 1 ⊗ J n × n ) + I q 1 n × q 1 n 2 J p 2 × q 1 n J p 3 × q 1 n 2 J p 1 n × p 2 J q 1 n × p 2 ( 2 ( J − I ) − A ( G 2 ) ) p 2 × p 2 2 J p 3 × p 2 2 J p 1 n × p 3 J q 1 n × p 3 2 J p 2 × p 3 ( 2 ( J − I ) − A ( G 3 ) ) p 3 × p 3
The proof follows from the same arguments as in Theorem 3.4 .
Theorem 4.5. Let G G G be an r r r -regular graph on p p p vertices and q q q edges, and H H H be a regular graph with the smallest eigenvalue at least − 2 -2 − 2 . For a fixed t ∈ N t\in\mathbb{N} t ∈ N , let T \mathcal{T} T be the family of all integer partitions T = ( t 1 , t 2 , … , t s ) T=(t_1,t_2,\ldots,t_s) T = ( t 1 , t 2 , … , t s ) of t t t into parts t i ≥ 3 t_i\ge3 t i ≥ 3 , ( 1 ≤ i ≤ s ) (1\le i\le s) ( 1 ≤ i ≤ s ) . For each such T ∈ T T\in\mathcal{T} T ∈ T , let C T = C t 1 ∪ C t 2 ∪ ⋯ ∪ C t s C_T=C_{t_1}\cup C_{t_2}\cup\cdots\cup C_{t_s} C T = C t 1 ∪ C t 2 ∪ ⋯ ∪ C t s be the corresponding disjoint union of cycles. Then,
(i) the graphs S G ( n , n ) ∨ ˙ ( C T ∪ H ) S_G(n,n)\dot{\vee}(C_T\cup H) S G ( n , n ) ∨ ˙ ( C T ∪ H ) , T ∈ T T\in\mathcal{T} T ∈ T form a family of D-equienergetic graphs,
(ii) the graphs S G ( n , n ) ∨ ( C T ∪ H ) S_G(n,n)\vee(C_T\cup H) S G ( n , n ) ∨ ( C T ∪ H ) , T ∈ T T\in\mathcal{T} T ∈ T form a family of D-equienergetic graphs.
Acknowledgement
The research of Anjitha Ashokan is supported by the University Grants Commission of India under the beneficiary code BININ05086971.
Declarations
On behalf of all authors, the corresponding author states that there is no conflict of interest.