1. Introduction
Let G be a simple undirected graph on n vertices with vertex set and edge set . The distance of v and u is defined to be the length of a shortest path from u to v. The maximum value in the set is called the eccentricity of u, denoted by . For , write (or just for short if there is no confusion) for the degree of u in G, and for the neighborhood of u.
For a molecular, if we let vertices represent the atoms and edges represent the bonds, then the resulting graph is called a molecular graph. So, a molecular graph could clearly reveal the corresponding molecular structure. Moreover, one could discover a molecule’s chemical properties by investigating its molecular graph’s combinatorial properties. A topological index for a molecular graph
G is a numerical quantity invariant under automorphisms of
G. Topological indices bridge chemical compounds’ physical, chemical, and thermodynamic parameters [
1]. Up to now, researchers have defined many topological indices and used them to model chemical, pharmaceutical, and other properties of molecules. Nowadays, some novel computational techniques for topological indices have been developed, such as cut method, extended cut method, and vertex cut method, see, for example [
2,
3,
4]. Such methods provide uniform way to deal with different topological indices. As one of the classic topological indices, the Wiener index is strongly related to many physical and chemical properties of molecular compounds (for the recent survey on the Wiener index see [
5]). For all unordered pairs of distinct vertices of
G, the summation of their distances is called the Wiener index of
G and is denoted by
, that is,
In 1994, Dobrynin and Kochetova [
6], and Gutman [
7] independently proposed a weighted version of the Wiener index as follows,
where
The value is also called the
degree distance of
G. It is interesting that, if
G is a tree, then
(see [
7]).
In [
7], another interesting index was also proposed, which is called the
Schultz index of the second kind and also called the Gutman index somewhere (see [
8] for example). It is defined to be
Also, if
G is a tree on
n vertices, then
(see [
7]).
In [
9], some extremal properties of the degree distance of graphs were reported. Dankelmann et al. [
10] presented an asymptotically sharp upper bound of degree distance of graphs with given order and diameter. In [
11], the authors determined the bicyclic graph with maximal degree distance. Ilić et al. [
12] calculated the degree distance for partial Hamming graphs. In [
13], Tomescu obtained the minimum degree distance of unicyclic and bicyclic graphs. The Gutman index of graphs attracts attention just recently, see, for example, [
14,
15,
16,
17]. The maximal and minimal Gutman index of bicyclic graphs were studied in [
18,
19]. In [
20], the authors presented an asymptotic upper bound for the Gutman index and also established the relation between the edge-Wiener index and the Gutman index of graphs.
A unicyclic graph is a connected graph obtained from a tree by adding an edge connecting its two vertices. Denote by the cycle of n vertices. Let G be a unicyclic graph containing the cycle . By deleting all edges in , we obtain some disjoint trees. Each of these trees contains exactly one vertex of , which is the root of such tree in G. These trees are called the branches of G. Let M be a subset of edges of G. If any pair of edges does not share a common vertex, then M is called a matching of G, and a vertex incident to some edge of M is said to be M-saturated. Particularly, if all vertices of G are M-saturated, then M is a perfect matching.
For integers
and
m,
, let
be the set of unicyclic graphs with
n vertices and matching number
m. Obviously, if
, then
G is the triangle. In the following we assume that
. Denote by
the graph obtained by connecting
new edges and
new vertices to a common vertex of the triangle
(see
Figure 1). Clearly,
.
By immediate calculations, we have
In this paper, we study the Gutman index of unicyclic graphs with given matching number and determine extremal graphs with the minimum Gutman index. On the one hand, we find that the graph minimizing the Gutman index among
plays an important role in dealing with
. So, we first deal with the special case of
. On the other hand, we use induction method to deal with
instead of computational methods, which is very different from the earlier papers on this topic. Let
be the graph obtained by attaching a pendant vertex to every vertex of a triangle,
the graph obtained from
by attaching one pendant vertex to a vertex of degree 3 in
, and let
be the graph obtained by attaching three pendant vertices to three consecutive vertices of
(see
Figure 2). In fact, we obtain the following results.
Theorem 1. Let , where .
- (i)
If , then with equality if and only if .
- (ii)
If , then with equality if and only if .
Theorem 2. Let , where .
- (i)
If , then with equality if and only if .
- (ii)
If , then with equality if and only if or for ; otherwise.
Note that, the extremal graphs for many indices, such as the spectral radius, the Wiener index, the Gutman index, and so on, are of high symmetry. It is interesting to investigate the inner relations between the symmetry of the graphs and their indices.
2. Main Results
Lemma 1 ([
21]).
Let , where , and let T be a branch of G with root r. If is a pendant vertex furthest from the root r with , then u is adjacent to a vertex of degree two. Lemma 2 ([
5]).
Let , where , and let . Then there is a maximum matching M and a pendant vertex u of G such that u is not M-saturated. Lemma 3. Let G be an n-vertex unicyclic graph with a pendant vertex u being adjacent to vertex v, and let w be a neighbor of v different from u. Thenwith equality if and only if ; for any . Moreover, if thenwith equality if and only if ; for any . Proof. From the definition, we have
with equality if and only if
;
for any
.
Similarly, we have
with equality if and only if
;
for any
.
We denote by
the graph obtained from
by adding
pendant vertices to a vertex of
. In [
18] it is obtained
It is well known (see [
22]) that
Let
be the unicyclic graph obtained from
by attaching a pendant vertex and
pendant vertices to
and
, respectively, where
. Suppose the neighbor of
with degree 1 is
w.
We can deduce from above that
□
Lemma 4. Suppose that . If or , then .
Proof. If , then it can be checked that = 175. We discuss according to the parity of k for next.
Case 1. If
k is even, then
, where
Then is positive in , negative in . Hence is increasing in and decreasing in . Thus, takes its minimal value at or .
So we obtain that for , and therefore .
It is easy to see that
for .
Case 2. If
k is odd, then
, where
Then is positive in , negative in . Hence is increasing in and decreasing in . Thus, takes its minimal value at or .
So we obtain that for , and therefore .
It is easy to see that
for .
Combining the above cases, we complete the proof. □
For integer , let be the set of graphs in containing a pendant vertex whose neighbor is of degree two. Let .
Recall that is the graph obtained by attaching three pendant vertices to three consecutive vertices of . It is easy to see that for .
Lemma 5. Let with . Then .
Proof. If , then the result follows easily. If , then by Lemma 1, implies that or G is a graph of maximum degree three obtained by attaching some pendant vertices to a cycle. If , .
Suppose that . Then G is a graph of maximum degree three obtained by attaching some pendant vertices to a cycle , where .
If , then every vertex on the cycle has degree three, and therefore any vertex on the cycle is adjacent to a unique pendant vertex. A direct computation shows that:
If is even, then
If is odd, then as above
If , then or since , by Lemma 4, for some , we have .
If , then G is the graph obtained from by attaching a pendant vertex. By direct computations, we have for .
In the following, if G is a graph in with a perfect matching M, then G contains a pendant vertex u whose neighbor v is of degree two in G, and assume w is the neighbor of v different from u. Obviously, . Since , we have . □
By immediate calculations, we could verify the following lemma.
Lemma 6. Among the graphs in , is the unique graph with minimum Gutman index 81; and is the only graph with the second minimum Gutman index 85.
Lemma 7. Let . Then with equality if and only if .
Proof. If
, then by Lemma 4,
. If
, then
. If
, then by Lemma 3
with equality if and only if
,
,
, i.e.,
.
If
, then
, and by Lemma 3,
The result follows. □
Lemma 8. Let . Then with equality if and only if .
Proof. If
, then by Lemma 4,
. If
. Then
. By Lemma 3
with equality if and only if
,
,
, i.e.,
. □
In the rest of the paper, we are going to present the proofs of the main results described in
Section 1.
Proof of Theorem 1.
The case is obvious since , , . The cases and follow from Lemmas 6 and 7, respectively.
Suppose that
. Let
. We prove the result by induction on
m. If
, then the result follows from Lemma 8. Suppose that
and the result holds for graphs in
. Let
. If
, then
G contains a pendant vertex
u whose neighbor
v is of degree two in
G, and assume
w is the neighbor of
v different from
u. By Lemma 4,
. If
, then
, and thus by Lemma 3 and the induction hypothesis, it is easily seen that
with equality if and only if
,
,
, i.e.,
. □
Recall that is the graph obtained from by attaching one pendant vertex to a vertex of degree 3 in .
Now, we can prove Theorem 2.
Proof of Theorem 2. The case follows from Lemma 6. Suppose that . Let .
For , we have . For with , we have either , bear in mind that , or ,
If with , then by Lemma 2, there exists a pendant vertex x and a maximum matching M such that x is not M-saturated in G, and thus . Let y be the unique neighbor of x. Since M contains one edge incident with y, and there are edges of G outside M, we have .
Case 1.. The result for
is obvious as in previous theorem. For
, it may be checked directly the five possibilities for
G to obtain that
has the minimum Gutman index 46. For
, it is known in [
23,
24] that
is the unique unicyclic graph on
n vertices with minimum Gutman index.
Case 2.. The result for
is obvious as in previous lemma. If
, then
. If
, then
, and by Lemma 3,
with equalities if and only if
and
, i.e.,
.
If
, then by Lemma 3 we have
with equalities if and only if
.
and
, i.e.,
. Thus, for
, we have
with equality if and only if
or
.
For
, we prove the result by induction on
n. If
, then
. By Lemma 3,
with equalities if and only if
,
and
, i.e.,
.
Suppose that
and the result holds for graphs in
. By Lemma 3 and the induction hypothesis,
with equalities if and only if
,
and
, i.e.,
.
Case 3.. The case
follows from Lemma 7. For
, we prove the result by induction on
n. If
, then
, and by Lemmas 7 and 3,
with equalities if and only if
,
and
, i.e.,
.
Suppose that
and the result holds for graphs in
. By Lemma 3 and the induction hypothesis,
with equalities if and only if
,
and
, i.e.,
.
Case 4.. We prove the result by induction on
n. If
, then the result follows from Theorem 1. Suppose that
and the result holds for graphs in
. Let
. By Lemma 3 and the induction hypothesis, it is easily seen that
with equality if and only if
,
,
, i.e.,
.
Combining the above cases, the result follows. □