Home 𝕮-inverse of graphs and mixed graphs
Article Open Access

𝕮-inverse of graphs and mixed graphs

  • Omar Alomari EMAIL logo , Mohammad Abudayah and Manal Ghanem
Published/Copyright: February 4, 2025

Abstract

This article introduces a generalization of the concept of inverse graphs applicable to both graphs and mixed graphs. Given a graph G with adjacency matrix A ( G ) , the inverse graph G 1 is defined such that its adjacency matrix is similar to the inverse of A ( G ) through a diagonal matrix with entries of ± 1 . While this diagonal matrix may or may not exist for graphs with nonsingular adjacency matrices, our study extends the concept to include mixed graphs as well. It has been proven that for certain unicyclic graphs, such a diagonal matrix does not exist. Motivated by this, we generalized the definition of inverse graphs to include mixed graphs, allowing us to find inverse mixed graphs for a class previously shown to lack one.

MSC 2010: 05C20; 05C50

1 Introduction

Let G = ( V ( G ) , E ( G ) ) be an unoriented simple graph, where V ( G ) = { v i } i = 1 n and E ( G ) = { v i v j : v i , v j V ( G ) } denote the vertex set and the edge set of G , respectively. A walk in G is a sequence of vertices, where every two consecutive vertices in the walk form an edge. The length of the walk is defined to be the number of edges in the walk (including repeated edges). A closed walk is a walk where it starts and ends with the same vertex. A cycle is a closed walk of different vertices. A matching of a graph G is a subset M of E ( G ) such that no two edges of M have a common vertex. A matching M is called perfect matching if for every v i V ( G ) , there is an edge of M incident to v i . A bipartite graph G is a graph where its vertices, V ( G ) , can be partitioned into two sets V 1 and V 2 , such that no two vertices of the same set form an edge. Obviously, a graph G is bipartite if and only if G has no odd cycle.

Let G = ( V ( G ) , E ( G ) ) be a graph and ξ : E ( G ) V ( G ) × V ( G ) E ( G ) be a function such that ξ ( v i v j ) { v i v j , ( v i , v j ) , ( v j , v i ) } . Then, a mixed graph is defined to be partially oriented graph G ξ = ( V ( G ) , E ξ ( G ) ) , where E ξ ( G ) = ξ ( E ( G ) ) , here G is called the underlying graph of G ξ . If ξ ( v i v j ) = ( v i , v j ) , then the edge ( v i , v j ) is called an arc from v i to v j , while if ξ ( v i v j ) = v i v j , then the edge v i v j is called digon.

Exploring the structural properties of graphs through associated matrices has long been a focal point of research in graph theory. Among the various matrices linked to a graph, the adjacency matrix stands out as one of the most prominent and extensively studied. Denoted as A ( G ) = [ a i j ] , the adjacency matrix of a graph G is a symmetric matrix of size n , where

(1) a i j = 1 if v i v j in E ( G ) , 0 otherwise .

In recent years, there has been a growing interest in the examination of matrices associated with mixed graphs. Among these, the complex unit gain graph [1] has emerged as a prominent subject of study, building upon the framework of three-colored digraphs introduced by Kalita [2]. Recent investigations and results can be found in previous studies [38].

Let T n denote the multiplicative group of all n th roots of unity. A T n -gain graph is defined as the triple G ξ , n ϕ = ( G ξ , T n , ϕ ) , consisting of a mixed graph G ξ = ( V ( G ) , E ξ ( G ) ) , the group T n and the gain function ϕ : E ( G ) T n such that for every digon e = v i v j of G ξ , ϕ ( e ) = 1 . The ϕ -Hermitian adjacency matrix of G ξ , n ϕ is defined by H ϕ = [ h i j ] , where

(2) h i j = ϕ ( v i v j ) if ξ ( v i v j ) = ( v i , v j ) , ϕ ( v i v j ) ¯ if ξ ( v i v j ) = ( v j , v i ) , 1 if ξ ( v i v j ) = v i v j , 0 otherwise .

Note here a complex unit gain subgraph H ξ , n ϕ of a complex unit gain graph G ξ , n ϕ is defined to be a subgraph of G together with the same orientation and complex unit gain functions but with restriction over the edges of the subgraph H of G . Obviously, G ξ , 1 ϕ is exactly the adjacency matrix of the underlying graph G . Furthermore, for a mixed graph G ξ , n ϕ , if ϕ ( E ξ ) { 1 , α , α ¯ } , then H ϕ will reduce to the α -Hermitian adjacency matrix which was defined in [9] as follows.

Let G ξ be a mixed graph and α be a unit complex number. Then, the α -Hermitian adjacency matrix of G ξ is the square matrix H α = [ h i j ] where

(3) h i j = α if ξ ( v i v j ) = ( v i , v j ) , α ¯ if ξ ( v i v j ) = ( v j , v i ) , 1 if ξ ( v i v j ) = v i v j , 0 otherwise .

Originating from a chemistry problem, the concept of inverse graphs was initially introduced by Godsil in [10]. A graph G is called nonsingular if its adjacency matrix lacks singularity. Moreover, G is called invertible if A 1 ( G ) bears a signature similarity to the adjacency matrix of another graph G . In other words, there exists a ± 1 diagonal matrix D such that D A 1 D forms the adjacency matrix of a mixed graph. Godsil [10] characterizes a subclass of bipartite graphs that possesses an inverse, while also posing the problem of characterizing bipartite graphs with inverses, which remains an open question enticing researchers. Kalita and Sarma [11] were the first to study the inverse of unicyclic three-colored digraphs. In their study, they explored the inverse of unicyclic three-colored digraphs. Additionally, Akbari and Kirkland [12] characterized when bipartite unicyclic graphs with unique perfect matching possess an inverse.

This study presents a generalization of Godsil’s classical definition and extends the findings of Kalita and Sarma. Moreover, we establish that bipartite unicyclic graphs lacking inverse graphs possess a mixed graph inverse (referred to as C -Inverse in this context). Furthermore, for a mixed graph G ξ , we construct an orientation for the complex unit gain graph G associated with H α 1 to obtain a γ -Hermitian adjacency matrix for G . This advancement holds promise in surmounting various challenges in this research line.

2 Preliminaries

The singularity of a graph’s adjacency matrix can vary; it may be singular or nonsingular. Research has illustrated the significant impact that the matching of a graph G has on the expansion of its determinant. The determinant expansion for various matrices associated with graphs, including the adjacency matrix of graphs G , the α -Hermitian adjacency matrix of mixed graphs G ξ , and the ϕ -Hermitian adjacency matrix of complex unit gain graphs G ξ , n ϕ , is provided in [13] and [14]. The subsequent definitions are necessary to introduce the determinant expansion.

Definition 1

Let G ξ be a mixed graph, then

  1. A mixed subgraph X ξ of G ξ is called elementary subgraph if each component of X is either edge or cycle.

  2. The rank and co-rank of an elementary mixed subgraph are defined by r ( X ξ ) = n c and s ( X ξ ) = m r ( X ξ ) , respectively, where n = V ( X ) , c is the number of components of X , and m = E ( X ) .

  3. If G has a unique perfect matching, then a path P ξ of G ξ is called m m -alternating path if the edges of P start and end with matching edges and its edges alternate between matching and unmatching edges.

  4. If P : v i 1 v i 2 v i 3 v i k is a path in G ξ , then the weight of P in the complex unit gain graph G ξ , n ϕ is defined by

    H ϕ ( P ) = j = 1 k 1 h i j i j + 1 .

We will now present a result similar to graph adjacency matrices (refer to [15], p. 44). The details can be located in [16].

Theorem 1

Let G ξ , n ϕ be a complex unit gain graph and H ϕ be its ϕ -Hermitian adjacency matrix. Then,

det ( H ϕ ) = X ( 1 ) r ( X ) 2 s ( X ) Re ( h ϕ ( C ) ) ,

where the sum ranges over spanning elementary subgraphs of G, the product is being taken over all cycles C in X and C is any mixed closed walk traversing C.

Let G be a unicyclic bipartite graph with unique perfect matching M and the cycle C . Then, one can easily check that the only elementary mixed graph is M itself. Therefore, using Theorem 1, we obtain

det ( H ϕ ) = ( 1 ) n 2 .

By employing Theorem 1, Abudayah et al. [17] provided a formula for the entries of the inverse of the α -Hermitian adjacency matrix of a bipartite mixed graph G ξ with a unique perfect matching in Theorem 8. Utilizing a similar technique to their proof, we extend this theorem as follows.

Theorem 2

Let G ξ , n ϕ be a bipartite mixed graph with unique perfect matching , H ϕ be its ϕ -Hermitian adjacency matrix, and

i j = { P i j : P i j is a n m m - a l t e r n a t i n g p a t h f r o m t h e v e r t e x i to t h e v e r t e x j in G } .

Then,

( H ϕ 1 ) i j = P i j i j ( 1 ) E ( P i j ) 1 2 h ϕ ( P i j ) if i j , 0 if i = j .

Definition 2

Let G be a unicyclic bipartite graph with unique perfect matching. Then, a matching edge is called a peg if it is incident to exactly one vertex of the cycle of G .

It is worth noting that in a bipartite unicyclic graph G with a unique perfect matching M , the number of vertices in the matching should be even. Additionally, if G contains more than two vertices, then there exists at most one mm-alternating path between any two vertices v i and v j of G . The proof of the upcoming theorem mirrors the proof of Corollary 3 in [17].

Theorem 3

Let G be a unicycle bipartite graph with unique perfect matching. If G has more than two pegs, and H ϕ is the ϕ -Hermitian adjacency matrix of the complex unit gain graph G ξ , n ϕ , then

[ H ϕ 1 ] i j = ( 1 ) E ( P i j ) 1 2 h ϕ ( P v i v j ) if P v i v j is t h e mm - alternating p a t h b e t w e e n v i and v j , 0 otherwise .

Akbari and Kirkland [12] characterized when bipartite unicyclic graphs with unique perfect matching has an inverse.

Theorem 4

Let G be a bipartite unicyclic graph having a unique perfect matching. Let the cycle be of length 2 m , and suppose that there are 2 k pegs. Then, G is invertible if and only if one of the following holds:

  1. k 2 and m k is even,

  2. k = 1 , m is even, and the vertices on the cycle incident with pegs are adjacent.

3 C -Inverse unicyclic bipartite graph with unique perfect matching with k 2 and m k is odd

Akbari and Kirkland [12] focused on identifying an inverse graph for the unicyclic bipartite graph G with a unique perfect matching M . The main outcome of this study is encapsulated in Theorem 4, where they classify the conditions under which G possesses an inverse graph. If conditions 1 and 2 of Theorem 4 are not satisfied, it implies that the graph G lacks an inverse graph. However, it may still possess an inverse mixed graph. To generalize the scope of graph inversion, consider the following definition.

Definition 3

Let G ξ be a mixed graph with α -Hermitian adjacency matrix H α . Then, G ξ is called C -invertible if there is ± α r diagonal matrix D and unit complex number γ , such that D H α 1 D * is γ -Hermitian adjacency matrix of some mixed graph ( I G ) η .

Let G be a unicyclic bipartite graph with a unique perfect matching M , adjacency matrix A , and cycle C . Let us assume that C has a length of 2 m and 2 k pegs, where k 2 . For simplicity, we denote the class of all unicyclic graphs with a unique perfect matching and k 2 by U . The subclass of U with the property m k is odd is denoted by U odd . If G belongs to U odd , Theorem 4 shows that there exists no ± 1 diagonal matrix D such that the property D A 1 D constitutes the adjacency matrix of a graph. The subsequent theorems explains the reasons behind the nonexistence of such a diagonal matrix.

Theorem 5

Let G be a graph, A be its adjacency matrix, D be ± α r diagonal matrix, where α is a primitive n th root of unity and G ξ , n ϕ is the complex unit gain graph corresponding to the matrix D A D * . Then, the weight of any cycle in G ξ , n ϕ is 1.

Proof

Suppose that C : v i 1 v i 2 v i 3 v i k v i 1 is a cycle G ξ , n ϕ and H ϕ = D A D * , where D = diag ( d 1 , d 2 , , d n ) . Then,

(4)□ h ϕ ( C ) = h ϕ ( i 1 i 2 ) h ϕ ( i 2 i 3 ) h ϕ ( i k i 1 ) = d i 1 d i 2 ¯ d i 2 d i 3 ¯ d i k d i 1 ¯ = 1 .

Corollary 1

Let G U , ( I G ) η be the mixed graph corresponding to the ( 1 ) -Hermitian adjacency matrix A 1 . Then, A 1 is not signable similar to 0 1 -matrix if and only if ( I G ) η contains a cycle of weight 1 with respect to the ( 1 ) -Hermitian adjacency matrix H 1 .

Example 1

In Figure 1, the graph G U with m = 3 and k = 2 . Note that, in the mixed graph ( I G ) η , the cycles C ( 1 ) : 0 , 9 , 7 , 5 , 0 and C ( 2 ) : 0 , 9 , 7 , 5 , 2 , 3 , 0 have the weight h 1 ( C ( 1 ) ) = h 1 ( C ( 2 ) ) = 1 . While the weight of the cycle C ( 3 ) : 5 , 2 , 3 , 0 , 5 has the weights h 1 ( C ( 3 ) ) = 1 , respectively. Therefore, there is no ± 1 diagonal matrix D such that D A 1 D is ( 0 1 ) -matrix (Figure 2).

Figure 1 
               A graph 
                     
                        
                        
                           G
                           ∈
                           U
                        
                        G\in {\mathfrak{U}}
                     
                  .
Figure 1

A graph G U .

Figure 2 
               The mixed graph 
                     
                        
                        
                           
                              
                                 
                                    (
                                    
                                       I
                                       G
                                    
                                    )
                                 
                              
                              
                                 η
                              
                           
                        
                        {\left(IG)}_{\eta }
                     
                   corresponds to the 
                     
                        
                        
                           
                              (
                              
                                 −
                                 1
                              
                              )
                           
                        
                        \left(-1)
                     
                  -Hermitian adjacency matrix 
                     
                        
                        
                           
                              
                                 A
                              
                              
                                 −
                                 1
                              
                           
                        
                        {A}^{-1}
                     
                  .
Figure 2

The mixed graph ( I G ) η corresponds to the ( 1 ) -Hermitian adjacency matrix A 1 .

In order to classify the cycles ( I G ) η for a graph G U according to their weights, we need the following elementary theorems.

Theorem 6

Let T be a tree and C be a cycle of ( I T ) η . Then, there are two vertices u and v of T such that V ( C ) V ( P u v ) , where P u v is the path from u to v in T .

Proof

Suppose not, then there are vertices r , s 1 , s 2 , and s 3 such that s 1 , s 2 , s 3 are vertices of the cycle C and for i j , V ( P r s i ) V ( P r s j ) = { r } , where P r s i and P r s j are paths in T . Choose such s j to be of maximum distance from r . Then, there are two vertices u 1 r s 1 and u 2 r s 2 of V ( P r s 1 ) and V ( P r s 2 ) , respectively, such that u 1 r s 1 and u 1 r s 2 are adjacent in C , two vertices u 2 r s 2 and u 1 r s 3 of V ( P r s 2 ) and V ( P r s 3 ) , respectively, such that u 2 r s 2 and u 1 r s 3 are adjacent in C , and two vertices u 2 r s 3 and u 3 r s 1 of V ( P r s 3 ) and V ( P r s 1 ) , respectively, such that u 2 r s 3 and u 3 r s 1 are adjacent in C . This contradicts the fact that r is matched by only one matching edge (Figure 3).□

Figure 3 
               The path from vertex 1 to vertex 8 is a source of a cycle in 
                     
                        
                        
                           
                              
                                 
                                    (
                                    
                                       I
                                       T
                                    
                                    )
                                 
                              
                              
                                 η
                              
                           
                        
                        {\left(IT)}_{\eta }
                     
                  .
Figure 3

The path from vertex 1 to vertex 8 is a source of a cycle in ( I T ) η .

Theorem 7

Let T be a tree, A be the adjacency matrix of T, and ( I T ) η be the mixed graph corresponding to ( 1 ) -Hermitian adjacency matrix A 1 . Then, the weight of any cycle C in ( I T ) η is h 1 ( C ) = 1 .

Proof

Let C be a cycle in ( I T ) η . As T is a tree, the edge v i v j belongs to C if and only if there exists an mm-alternating path between v i and v j in T . Consequently, there exists an mm-alternating path between any two consecutive vertices of C in T . Utilizing Theorem 2, the weight of cycle C in I T η equals ( 1 ) a , where a = u v E ( C ) n u v , and n u v denotes the number of non-matching edges between u and v in T . Given that T is a tree and mm-alternating paths start and end at the same vertex u of T , each non-matching edge between the vertices of C in T is counted an even number of times, as per Theorem 6. This implies that a is even.□

In a mixed graph G ξ , a cycle C is called self-balanced if the weight of C equals one regardless of the value of α . Equivalently, C is self-balanced if and only if (according to fixed direction) the number of forward arcs and the number of backward arcs are equal.

Theorem 8

Let T be a tree graph with unique perfect matching and ( I T ) η be the mixed graph corresponding to ( 1 ) -Hermitian adjacency matrix A 1 . Then, there is : E ( I T ) V ( I T ) × V ( I T ) E ( I T ) such that:

  1. If ( A 1 ) u v = 1 , then ( u v ) = u v .

  2. For any cycle C in ( I T ) , C is self-balanced.

Proof

Fix a vertex v of T , and define ξ v : E ( T ) V ( T ) × V ( T ) E ( T ) as follows: For every leaf u in T , starting from v , orient the non-matching edges of the path between v and u to be opposite to each other and fix the matching edges as digons (Figure 4). Now, for any unit complex number α , suppose that H α is the α -Hermitian adjacency matrix of T ξ . Then, using Lemma 3.2 in [1], there is a diagonal matrix D with α power entries, such that D A D * = H α . Using Theorem 2, one can easily see that the entries of matrix H α 1 are from the set { 0 , ± 1 , ± α , ± α ¯ } . Furthermore, since every tree graph with perfect matching is invertible [18] and using Theorem 3, there is a ± 1 - diagonal matrix S with the property S H α 1 S is Hermitian adjacency matrix of a mixed graph I T η for some orientation function η . Finally, using Theorem 5, ( I T ) η is α -balanced. This is true for any value of α , and thus every cycle of ( I T ) η is self-balanced.□

Figure 4 
               (a) Tree graph 
                     
                        
                        
                           T
                        
                        T
                     
                  , (b) the tree mixed graph 
                     
                        
                        
                           
                              
                                 T
                              
                              
                                 ξ
                              
                           
                        
                        {T}_{\xi }
                     
                  , and (c) the mixed graph 
                     
                        
                        
                           
                              
                                 
                                    (
                                    
                                       I
                                       T
                                    
                                    )
                                 
                              
                              
                                 η
                              
                           
                        
                        {\left(IT)}_{\eta }
                     
                  .
Figure 4

(a) Tree graph T , (b) the tree mixed graph T ξ , and (c) the mixed graph ( I T ) η .

For a graph G U , the following theorem classifies the cycles in ( I G ) η , which has the weight h 1 ( C ) = 1 .

Theorem 9

Let G U . Suppose that { u i u i } i = 1 2 k are the pegs of the G -cycle C, where u i V ( C ) . Then,

  1. For any cycle C of ( I G ) η with the property { u i } i = 1 2 k V ( C ) , C has the weight h 1 ( C ) = 1 .

  2. The cycle C induced by the vertices { u i } i = 1 2 k in ( I G ) η has the weight h 1 ( C ) = ( 1 ) m k .

Proof

(1) Suppose that C is a cycle in ( I G ) η and u r is not a vertex of C . Observing that u r u r is a peg of the cycle C in G , u r is incident to two non-matching edges of C say u r t and u r s . Obviously, T = G u r t is a tree with perfect matching. Furthermore, since v i v j is an edge of C if and only if there is an mm-alternating path between v i and v j in G , we have that C is a cycle in ( I T ) η . Using Theorem 7 we obtain h 1 ( C ) = 1 .

(2) Obviously, { u i } i = 1 2 k induces a cycle C in ( I G ) η . Also, h 1 ( u i u i + 1 ) = ( 1 ) a i , where a i is the number of non-matching edges in the mm-alternating paths between u i and u ( i + 1 ) ( m o d ( 2 k ) ) . Therefore,

(5) h 1 ( C ) = h 1 ( u 1 u 2 ) h 1 ( u 2 u 3 ) h 1 ( u k u 1 ) = ( 1 ) i = 1 2 k a i = ( 1 ) c ,

where c is the number of non-matching edges in the cycle C of G . Observing that C has 2 k pegs, the number of vertices incident to a matching edge e E ( C ) is 2 m 2 k . Therefore, the number of non-matching edges of C is 2 m 2 m 2 k 2 = m + k .□

Corollary 2

If G U odd and A its adjacency matrix, then there is no ± 1 diagonal matrix D with the property D A 1 D are { 0 , 1 } -matrix.

Corollary 3

Let G U odd and suppose that { u i u i } i = 1 2 k are the pegs of the G -cycle C, where u i V ( C ) . Then, a cycle C in ( I G ) η has the weight h 1 ( C ) = 1 if and only if { u i } i = 1 2 k V ( C ) .

Proof

Let C be a cycle in ( I G ) η with the property { u i } i = 1 2 k V ( C ) . Then, observing that { u i } i = 1 2 k induces a cordless cycles C m in ( I G ) η , for each i , { u i u ( i + 1 ) ( m o d ( 2 k ) ) } is either a chord in C or an edge in C . Therefore, C { C m } can be partitioned into cycles { C s } s = 1 j each of which contains exactly one edge of C m (Figure 5). Now, suppose that C r : { u i u ( i + 1 ) ( m o d ( 2 k ) ) v 1 v 2 u i } is one of these cycles, then

h 1 ( C r ) = h 1 ( u i u ( i + 1 ) ( m o d ( 2 k ) ) ) h 1 ( v 1 v 2 u i ) .

Using Theorem 9, h 1 ( C r ) = 1 . Therefore, h 1 ( u i u ( i + 1 ) ( m o d ( 2 k ) ) ) h 1 ( v 1 v 2 u i ) = 1 , and thus h 1 ( C ) = h ( C m ) = 1 . Finally, using Theorem 9, we obtain the result.□

Figure 5 
               The cordless cycle 
                     
                        
                        
                           C
                           ′
                        
                        C^{\prime} 
                     
                  , 
                     
                        
                        
                           
                              
                                 C
                              
                              
                                 m
                              
                           
                        
                        {C}_{m}
                     
                   and the cycles constructed by 
                     
                        
                        
                           
                              {
                              
                                 
                                    
                                       u
                                    
                                    
                                       i
                                    
                                    
                                       ′
                                    
                                 
                                 
                                    
                                       u
                                    
                                    
                                       
                                          (
                                          
                                             i
                                             +
                                             1
                                          
                                          )
                                       
                                       
                                          (
                                          
                                             m
                                             o
                                             d
                                             
                                                (
                                                
                                                   2
                                                   k
                                                
                                                )
                                             
                                          
                                          )
                                       
                                    
                                    
                                       ′
                                    
                                 
                              
                              }
                           
                        
                        \left\{{u}_{i}^{^{\prime} }{u}_{\left(i+1)\left(mod\left(2k))}^{^{\prime} }\right\}
                     
                  .
Figure 5

The cordless cycle C , C m and the cycles constructed by { u i u ( i + 1 ) ( m o d ( 2 k ) ) } .

The following theorem completes Theorem 12 in [12].

Theorem 10

Let G U odd with 2 k pegs. Then, G is C -invertible.

4 Conclusion

In this study, we generalize the concept of inverse graphs to accommodate mixed graphs. We prove that all graphs lacking inverse graph, as described in Theorem 10-a [1], possess inverse mixed graph. This opens the door for researchers to address several important problems in this area, particularly the one proposed by Godsil [10].

Acknowledgement

The authors wish to acknowledge the support by the Deanship of Scientific Research at German Jordanian University.

  1. Funding information: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

  2. Author contributions: All authors contributed equally.

  3. Conflict of interest: The authors state no conflicts of interest.

  4. Data availability statement: No data were used to support this study.

References

[1] N. Reff, Spectral properties of complex unit gain graphs, Linear Algebra Appl. 436 (2012), no. 9, 3165–3176. 10.1016/j.laa.2011.10.021Search in Google Scholar

[2] D. Kalita, On colored digraphs with exactly one nonsingular cycle, Electron. J. Linear Algebra 23 (2012), 397–421. 10.13001/1081-3810.1529Search in Google Scholar

[3] M. Cavaleri and A. Donno, On cospectrality of gain graphs, Spec. Matrices 10 (2022), no. 1, 343–365. 10.1515/spma-2022-0169Search in Google Scholar

[4] A. Alazemi, M. Andelic, F. Belardo, M. Brunetti, and C. M. da Fonseca, Line and subdivision graphs determined by T4-gain graphs, Mathematics 7 (2019), no. 10, 926. 10.3390/math7100926Search in Google Scholar

[5] O. Alomari, M. Abudayah, and M. Ghanem, On the cospectrality of Hermitian adjacency matrices of mixed graphs, Eur. J. Pure Appl. Math. 15 (2022), no. 3, 1090–1097. 10.29020/nybg.ejpam.v15i3.4422Search in Google Scholar

[6] M. Abudayah, O. Alomari, and T. Sander, Incidence matrices and line graphs of mixed graphs, Spec. Matrices 11 (2023), no. 1, 20220176. 10.1515/spma-2022-0176Search in Google Scholar

[7] W. Boulahmar, M. Ghanem, and M. Abudayah, γ-Inverse graph of some mixed graphs, Spec. Matrices 12 (2024), no. 1, 20240003. 10.1515/spma-2024-0003Search in Google Scholar

[8] N. Reff, Oriented gain graphs, line graphs and eigenvalues, Linear Algebra Appl. 506 (2016), 316–328. 10.1016/j.laa.2016.05.040Search in Google Scholar

[9] B. Mohar, A new kind of Hermitian matrices for digraphs, Linear Algebra Appl. 584 (2020), 343–352. 10.1016/j.laa.2019.09.024Search in Google Scholar

[10] C. D. Godsil, Inverses of trees, Combinatorica 5 (1985), 33–39. 10.1007/BF02579440Search in Google Scholar

[11] D. Kalita and K. Sarma, On the inverse of unicyclic 3-coloured digraphs, Linear Multilinear Algebra 70 (2022), no. 21, 6223–6239. 10.1080/03081087.2021.1948956Search in Google Scholar

[12] S. Akbari and S. J. Kirkland, On unimodular graphs, Linear Algebra Appl. 421 (2007), no. 1, 3–15. 10.1016/j.laa.2006.10.017Search in Google Scholar

[13] M. Abudayah, O. Alomari, and T. Sander, Hermitian adjacency matrices of mixed graphs, Eur. J. Pure Appl. Math. 15 (2022), no. 3, 841–855. 10.29020/nybg.ejpam.v15i3.4448Search in Google Scholar

[14] A. Samanta and M. R. Kannan, On the Spectrum of Complex Unit Gain Graph, 2019, arXiv: http://arXiv.org/abs/arXiv:1908.10668. Search in Google Scholar

[15] N. Biggs, Algebraic Graph Theory, 2nd edn., Cambridge University Press, Cambridge, 1994. Search in Google Scholar

[16] R. Mehatari, M. R. Kannan, and A. Samanta, On the adjacency matrix of a complex unit gain graph, Linear Multilinear Algebra 70 (2022), no. 9, 1798–1813. 10.1080/03081087.2020.1776672Search in Google Scholar

[17] M. Abudayah, O. Alomari, and O. AbuGhneim, Inverse of α-Hermitian adjacency matrix of a unicyclic bipartite graph, J. Combin. Math. Combin. Comput. 119 (2024), 63–73. 10.61091/jcmcc119-07Search in Google Scholar

[18] M. Abudayah, O. Alomari, and O. AbuGhneim, Inverse of Hermitian adjacency matrix of a mixed graph, Appl. Math. Inf. Sci. 16 (2022), 823–828. 10.18576/amis/160516Search in Google Scholar

Received: 2024-02-17
Revised: 2024-10-04
Accepted: 2024-11-21
Published Online: 2025-02-04

© 2025 the author(s), published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Special Issue on Contemporary Developments in Graphs Defined on Algebraic Structures
  2. Forbidden subgraphs of TI-power graphs of finite groups
  3. Finite group with some c#-normal and S-quasinormally embedded subgroups
  4. Classifying cubic symmetric graphs of order 88p and 88p 2
  5. Simplicial complexes defined on groups
  6. Two-sided zero-divisor graphs of orientation-preserving and order-decreasing transformation semigroups
  7. Further results on permanents of Laplacian matrices of trees
  8. Research Articles
  9. Dynamics of particulate emissions in the presence of autonomous vehicles
  10. The regularity of solutions to the Lp Gauss image problem
  11. Exploring homotopy with hyperspherical tracking to find complex roots with application to electrical circuits
  12. The ill-posedness of the (non-)periodic traveling wave solution for the deformed continuous Heisenberg spin equation
  13. Some results on value distribution concerning Hayman's alternative
  14. 𝕮-inverse of graphs and mixed graphs
  15. A note on the global existence and boundedness of an N-dimensional parabolic-elliptic predator-prey system with indirect pursuit-evasion interaction
  16. On a question of permutation groups acting on the power set
  17. Chebyshev polynomials of the first kind and the univariate Lommel function: Integral representations
  18. Blow-up of solutions for Euler-Bernoulli equation with nonlinear time delay
  19. Spectrum boundary domination of semiregularities in Banach algebras
  20. Statistical inference and data analysis of the record-based transmuted Burr X model
  21. A modified predictor–corrector scheme with graded mesh for numerical solutions of nonlinear Ψ-caputo fractional-order systems
  22. Dynamical properties of two-diffusion SIR epidemic model with Markovian switching
  23. Classes of modules closed under projective covers
  24. On the dimension of the algebraic sum of subspaces
  25. Periodic or homoclinic orbit bifurcated from a heteroclinic loop for high-dimensional systems
  26. On tangent bundles of Walker four-manifolds
  27. Regularity of weak solutions to the 3D stationary tropical climate model
  28. A new result for entire functions and their shifts with two shared values
  29. Freely quasiconformal and locally weakly quasisymmetric mappings in metric spaces
  30. On the spectral radius and energy of the degree distance matrix of a connected graph
  31. Solving the quartic by conics
  32. A topology related to implication and upsets on a bounded BCK-algebra
  33. On a subclass of multivalent functions defined by generalized multiplier transformation
  34. Local minimizers for the NLS equation with localized nonlinearity on noncompact metric graphs
  35. Approximate multi-Cauchy mappings on certain groupoids
  36. Multiple solutions for a class of fourth-order elliptic equations with critical growth
  37. A note on weighted measure-theoretic pressure
  38. Majorization-type inequalities for (m, M, ψ)-convex functions with applications
  39. Recurrence for probabilistic extension of Dowling polynomials
  40. A generalized fixed-point theorem for set-valued mappings in b-metric spaces
  41. Unraveling chaos: A topological analysis of simplicial homology groups and their foldings
  42. Global existence and blow-up of solutions to pseudo-parabolic equation for Baouendi-Grushin operator
  43. A characterization of the translational hull of a weakly type B semigroup with E-properties
  44. Some new bounds on resolvent energy of a graph
  45. Carmichael numbers composed of Piatetski-Shapiro primes in Beatty sequences
  46. The number of rational points of some classes of algebraic varieties over finite fields
  47. Singular direction of meromorphic functions with finite logarithmic order
  48. Pullback attractors for a class of second-order delay evolution equations with dispersive and dissipative terms on unbounded domain
  49. Eigenfunctions on an infinite Schrödinger network
  50. Boundedness of fractional sublinear operators on weighted grand Herz-Morrey spaces with variable exponents
  51. On SI2-convergence in T0-spaces
  52. Bubbles clustered inside for almost-critical problems
  53. Classification and irreducibility of a class of integer polynomials
  54. Existence and multiplicity of positive solutions for multiparameter periodic systems
  55. Averaging method in optimal control problems for integro-differential equations
  56. On superstability of derivations in Banach algebras
  57. Investigating the modified UO-iteration process in Banach spaces by a digraph
  58. The evaluation of a definite integral by the method of brackets illustrating its flexibility
  59. Existence of positive periodic solutions for evolution equations with delay in ordered Banach spaces
  60. Tilings, sub-tilings, and spectral sets on p-adic space
  61. The higher mapping cone axiom
  62. Continuity and essential norm of operators defined by infinite tridiagonal matrices in weighted Orlicz and l spaces
  63. A family of commuting contraction semigroups on l 1 ( N ) and l ( N )
  64. Pullback attractor of the 2D non-autonomous magneto-micropolar fluid equations
  65. Maximal function and generalized fractional integral operators on the weighted Orlicz-Lorentz-Morrey spaces
  66. On a nonlinear boundary value problems with impulse action
  67. Normalized ground-states for the Sobolev critical Kirchhoff equation with at least mass critical growth
  68. Decompositions of the extended Selberg class functions
  69. Subharmonic functions and associated measures in ℝn
  70. Some new Fejér type inequalities for (h, g; α - m)-convex functions
  71. The robust isolated calmness of spectral norm regularized convex matrix optimization problems
  72. Multiple positive solutions to a p-Kirchhoff equation with logarithmic terms and concave terms
  73. Joint approximation of analytic functions by the shifts of Hurwitz zeta-functions in short intervals
  74. Green's graphs of a semigroup
  75. Some new Hermite-Hadamard type inequalities for product of strongly h-convex functions on ellipsoids and balls
  76. Infinitely many solutions for a class of Kirchhoff-type equations
  77. On an uncertainty principle for small index subgroups of finite fields
  78. On a generalization of I-regularity
  79. Spin(8, ℂ)-Higgs bundles fixed points through spectral data
  80. Coloring the vertices of a graph with mutual-visibility property
  81. Embedding of lattices and K3-covers of an enriques surface
  82. Algorithm and linear convergence of the H-spectral radius of weakly irreducible quasi-positive tensors
  83. q-Stirling sequence spaces associated with q-Bell numbers
  84. Multiple G-Stratonovich integral in G-expectation space
Downloaded on 5.10.2025 from https://www.degruyterbrill.com/document/doi/10.1515/math-2024-0104/html
Scroll to top button