1. Introduction
All graphs considered in this paper are undirected, finite and simple. Let
be a graph with n vertices and m edges. For convenience, the path, cycle and complete graph of order n are denoted by
,
and
, respectively. Let C be a set, and the number of elements in C is denoted by
, let
denote the number of cycles of length i.
The adjacency matrix of G is denoted by
. The polynomial
is called the characteristic polynomial of a graph G, where I is the identity matrix of order n. The spectrum of G consists of the eigenvalues together with their multiplicities of
. The spectral radius of G, denoted by
, is the maximum eigenvalue of graph G. The nullity of G, denoted by
, is the multiplicity of zeros in the spectrum of G. Let
be the rank of
. Obviously,
. Two graphs G and H are said to be cospectral (denoted by
) if they share the same spectrum. A graph G is said to be determined by its spectrum (DS for short) if for any graph H,
implies that H is isomorphic to G.
The energy of G first is defined by Gutman in 1978 as the sum of the absolute values of its eigenvalues. That is,
The theory of graph energy is well developed nowadays; its details can be found in the book [1] and reviews [2] [3] .
Definition 1.1. ( [4] ) Given a graph G with the set of vertices
and a vector of positive integers
, denote by
(Gom for short) the graph obtained from G by replacing each vertex
of G with an independent set of
vertices
and joining
with
if and only if
and
are adjacent in G. The resulting graph Gom is said to be obtained from G by multiplication of vertices. For graph
, we denote by
the class of all graphs that can be obtained from one of the graphs in
by multiplication of vertices.
By Definition 2.1, Chang et al. [4] characterized all connected graphs with rank 4. That is, if G is a connected graph with rank 4, then
, the resulting graph, see Figure 1. Wu et al. [5] studied further the spectral properties of graphs with rank 4. They computed the characteristic polynomials of all graphs with rank 4. And they showed that some graphs with rank 4 are determined by their spectra. In particular, they proposed a problem: Which graphs with rank 4 are determined by their spectra? Recently, Monsalve and Rada [6] characterized spectral radius of all connected graphs with rank 4. A natural problem is: How to characterize the spectral radius of all graphs with rank 4?
In this paper, we intend to solve these two problems. Preliminaries are presented in Section 2. And we give the expressions of the spectral radius and energy of all graphs with rank 4 in Section 3. In Section 4, we consider which graphs
Figure 1.
.
with rank 4 are DS. More precisely, we prove that two classes of graphs with rank 4 are DS. And some cospectral graphs with rank 4 are presented.
2. Some Lemmas
Several lemmas are of importance to the description and proof of our results later, and we list them below.
By the properties of vertex multiplication, Wu et al. [5] computed the characteristic polynomials with rank 4 as follows.
Lemma 2.1. ( [5] ) Let G be a simple graph on n vertices and
. Then
if an only if
, where the graphs
are depicted in Figure 1.
Lemma 2.2. ( [7] ) Let G be a graph. For the adjacency matrix the following can be deduced from the spectrum:
(i) The number of vertices.
(ii) The number of edges.
(iii) Whether G is regular.
(iv) Whether G is regular with any fixed girth.
(v) The number of closed walk of any length.
(vi) Whether G is bipartite.
Lemma 2.3. ( [8] ) Let G be a simple bipartite graph with e edges. Then
with equality if G is a disjoint union of a complete bipartite graph and isolated vertices.
Lemma 2.4. ( [5] ) Suppose that
, where
is depicted in Figure 1,
,
,
,
,
,
,
,
. Then each of the following holds.
(i)
.
(ii)
.
(iii)
.
(iv)
.
(v)
.
(vi)
.
(vii)
.
(viii)
.
(ix)
.
3. The Spectral Radii and Energies of Graphs with Rank 4
In this section, we give the spectral radii and energies of graphs with rank 4. All the notation in this paragraph is followed in Lemma 2.4.
Theorem 3.1. Let
. Then the spectral radius of graph G as follows:
(i) If
. Then
(ii) If
and let
. Then
where
where
(iii) If
and let
,
,
. Then
where
where
(iv) If
and let
,
,
. Then
where
where
(v) If
and let
,
,
. Then
where
where
(vi) If
and let
,
,
. Then
where
where
(vii) If
and let
,
,
. Then
where
where
(viii) If
. Then
(ix) If
. Then
Proof. Here we only consider the cases
,
. The proof of other cases is quite similar to
and is thus omitted.
Let
. By Lemma 2.3, directly yields
.
Let
. By Theorem 2.4 (ii), there exist 4 nonzero eigenvalues and all other eigenvalues are 0. So we only need to input polynomial
in maple 13.0, we can get the nonzero eigenvalues of the graph G as follows.
where
Due to
,
and
, we can obviously get
is the spectral radius of graph G. Let
,
,
, due to
Then we have
So we get
It is consistent with the spectral radius obtained as above.
This completes the proof. □
Example 1. Solve the spectral radius of graph
.
By employing maple 13.0 to calculate, we can get that 1.5808, −5.3747, −9.5359, 13.3297 are the nonzero eigenvalues of the graph G. By comparison, it is obvious that 13.3297 is the spectral radius of the graph G.
Theorem 3.2. Let
. Then the energy of graph G as follows, where the notations is defined as same as above Theorem.
(i) If
. Then
(ii) If
. Then
(iii) If
.Then
(iv) If
. Then
(v) If
. Then
(vi) If
. Then
(vii) If
. Then
(viii) If
. Then
(ix) If
. Then
Proof. Here we only consider the cases
. The proof of other cases is quite similar to
and is thus omitted.
The proof of Theorem 3.2 follows from Theorem 3.1. So we have
Due to
,
, we get
It is consistent with the energy obtained as above.
This completes the proof. □
4. The Spectral Characterization of Graphs with Rank 4
In this section, we will investigate which graph
is DS and find some cospectral graphs.
Theorem 4.1. Let
, where
in
. Then G is DS.
Proof. Suppose that G has
vertices. Checking G, we note that it only contains one triangle. This implies, by Lemma 2.2 (v), that if graph H is cospectral with G, then H must contain one triangle. By Lemma 2.1 and Lemma 2.4,
(here
in
),
(here
in
) and
(here
in
) contains one triangle, respectively. It has been proved that
(here
in
) is DS. In the following we consider two cases.
Case 1. Assume that G and
are cospectral and let
,
. Therefore, G and
have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iv) and (iii), we have
Solving the equation system as above, we obtain that
, which implies
and
. By
and
, we can obtain that
. Taking
into
, we obtain that
. Solving this equation, we obtain that
or
. However, by
and
,
, we obtain that
or
, which in contradict with
or
. Hence G and
are not cospectral.
Case 2. Assume that G and
are cospectral. Therefore, G and
have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iv) and (v), we obtain that
Solving the equation system as above, we have
By
, we obtain that
. Taking
into
, we obtain that
. Supposing the roots of the equation as above are
, we have
By the definition of
, one has
, which implies that
,
. By
, we know that
are nonzero and have the same sign. However, by
, we know that
. This contradicts the fact
. Thus, G and
are not cospectral.
Next, assume that
and
are cospectral and
. Therefore, they have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iv), we get that
Solving the equation system as above, we obtain that
,
. By
, we obtain that
. Taking
into
, we have
. Solving this equation, we obtain that
or
. If
,
, then we have
, a contradiction; If
,
, then we have
, a contradiction. Thus,
and
are not cospectral.
From the argument above, we obtain that G is DS. □
Theorem 4.2. Let
, where
in
. Then G is DS if and only if
or
has no positive integer solution.
Proof. Suppose that G has
vertices. By Lemma 2.2 (v), we know if graph H is cospectral with graph G, then H must contain one triangle. By Lemma 2.1 and Theorem 2.4,
(here
),
(here
) and
(here
) contain one triangle, respectively. It has been proved that
(here
) is DS. In the following we consider two cases.
Case 1. Assume that G and
are cospectral and let
,
. Therefore, G and
have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iii) and (iv), we have
Solving the equation system as above, we obtain that
, which implies that
and
. By
and
, we can obtain that
. Taking
into
, we obtain that
. Solving the equation as above, we obtain that
or
. However, by
,
,
, we obtain that
or
, which in contradict with
or
. Thus, G and
are not cospectral.
Case 2. Assume that G and
are cospectral. Therefore, G and
have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iii) and (v), we have
Solving the equation system as above, we have
By
, we obtain that
. Taking
into
, we obtain that
. If
and
has positive integer solution are satisfied at the same time, then G and
are cospectral. On the contrary, G and
are not cospectral.
Then, assume that
and
are cospectral and let
. Therefore, they have the same vertices and coefficients of their characteristic polynomials. By Lemma 2.4 (iii), we have
Solving the equation system as above, we obtain that
,
. By
, we have
. Taking
into
, we have
. Solving the equation as above, we obtain that
or
. If
,
. Then we have
, which in contradict with
; When
,
. Then we have
, which in contradict with
. Thus,
and
are not cospectral.
In conclusion, we obtain that G is DS if and only if
or
has no positive integer solution.
□
Corollary 4.3. Let
where
in
and
where
in
. They are cospectral if and only if
and
has positive integer solution, where the graphs
are depicted in Figure 2.
Proof. By Theorem 4.2, we can obtain it obviously. □
5. Conclusion
In this paper, we give the expressions of the spectral radius and energy of all graphs with rank 4. At the same time, we investigate some graph
is DS and find some cospectral graphs.
Acknowledgements
This work is supported by the planning project of Qinghai Minzu University (No. 2021XJXS18), the innovation project of Qinghai Minzu University (No. 07M2022001).