1. Introduction
Throughout this paper, all graphs are assumed to be finite and undirected.
For a graph , we use , , and to denote the vertex set, edge set, arc set and full automorphism group of the graph , respectively. A graph is said to be arc-transitive if the full automorphism group acts transitively on . We use to denote the valency of the , and we say is a cubic, tetravalent, pentavalent or hexavalent graph, meaning , 4, 5 or 6.
Let
G be a finite group with identity element 1 and
S (say Cayley subset) a subset of G such that
and
. Define the
Cayley graph , that is, the Cayley graph of
G with respect to the Cayley subset
S as the graph with vertex set
G such that
are adjacent if and only if
. It is easy to see that the valency of
is
. As we all know,
is connected if and only if
. On the other hand, letting
be the right regular representation of
G and letting
be the full automorphism group of
, there are clearly
, and
acts transitively on the vertices of
. Then, the graph
is vertex-transitive, and
G (or
) can be viewed as a regular subgroup of
. Conversely, a connected graph
is isomorphic to a Cayley graph of a group
G if and only if the full automorphism group
contains a subgroup which acts regularly on
and the subgroup is isomorphic to
G (see [
1]). A Cayley graph
is said to be a
normal Cayley graph if the base group
G is normal in
; otherwise,
is said to be a
nonnormal Cayley graph (see [
2]).
The study about Cayley graphs on finite non-abelian simple groups has always attracted much attention because of Cayley graphs with high levels of symmetry; for example, vertex-transitivity, edge-transitivity and arc-transitivity are widely used in the design of interconnection networks. For more detailed applications, we recommend that readers refer to [
3,
4]. Let
G be a finite non-abelian simple group, and let
be a connected arc-transitive Cayley graph on
G. The main motivation for classifying 2-arc-transitive nonnormal Cayley graphs comes from the fact that Fang, Ma and Wang [
5] proved all but finitely that many locally primitive Cayley graphs of valency
or a prime number of the finite non-abelian simple groups are normal. In [
5] (Problem 1.2), they proposed the following problem: classify nonnormal locally primitive Cayley graphs (note that 2-arc-transitive graphs must be locally primitive) of finite simple groups with valency
or a prime number. To solve this problem, we should study each valency
or a prime number. In the case where
is a cubic graph (3-valent), Li [
6] proved that
must be normal, except for seven exceptions. On the basis of Li’s result, Xu et al. [
7,
8] proved that
must be normal, except for two exceptions on
. In the case where
is a tetravalent graph (4-valent), Fang et al. in [
9] proved that most of such
are normal, except for Cayley graphs on a list of
G. Further, Fang et al. in [
10] proved that
are normal when
is 2-transitive, except for two graphs on
. In the case where
is a pentavalent graph (5-valent), Zhou and Feng [
11] proved that all 1-transitive Cayley
of simple groups are normal. Ling and Lou in [
12] gave an example of a 2-transitive pentavalent nonnormal Cayley graph on
. Therefore, the next natural problem is to study the case of the 6-valent. However, there are no known nonnormal examples of hexavalent 2-arc-transitive Cayley graphs on finite simple groups.
The aim of this paper is to construct a nonnormal example of a connected 2-arc transitive hexavalent Cayley graph on a finite non-abelian simple group. Our main result is the following theorem.
Theorem 1. There exists a nonnormal example of a connected 2-arc-transitive hexavalent Cayley graph on the alternating group , and the full automorphism group of this graph is isomorphic to the alternating group .
2. Preliminaries
In this section, we give some necessary preliminary results which are used in later discussions.
Let
G be a finite group and let
H be a subgroup of
G. Then we have the following result (see [
13] (Ch. I, 1.4)).
Lemma 1. Let G be a group and let H be a subgroup of G. Let be the normalizer of H in G, and let be the centralizer of H in G. Then, is isomorphic to a subgroup of the automorphism group of H.
We next introduce the definition of a Sabidussi coset graph. Let
G be a group,
such that
, and let
H be a core-free subgroup of
G. Define the
Sabidussi coset graph of
G with respect to the core-free subgroup
H as the graph with vertex set
(the set of cosets of
H in
G) such that
and
are adjacent if and only if
. The following lemma follows from [
14], and it can be easily proved by the definition of the coset graphs (see [
15] (Theorem 3) for example).
Lemma 2. Let G be a group, such that , and let H be a core-free subgroup of G. Let be a Sabidussi coset graph of G with respect to H. Then, Γ is G-arc-transitive and the following holds:
- (1)
The valency of the graph Γ is equal to .
- (2)
Γ is a connected graph if and only if .
- (3)
If G contains a subgroup R is regular on , then , where .
Conversely, if Σ is an X-arc-transitive graph, then Σ is isomorphic to a Sabidussi coset graph , where is a 2-element such that , and , .
Proof. Let be an X-arc-transitive graph. Let be a vertex of and . Since is X-arc-transitive, there is g such that . For each , define . Then we can verify that is a graph isomorphic from to . Since is undirected, we have . Hence, . Thus, we can choose a 2-element g satisfying . □
Let
and
be two integers. We denote by the
-group the finite group of the order
. Following the definition of relevant objects in [
16] (Theorem 3.1), we have the following lemma, which is about the stabilizers of arc-transitive hexavalent graphs. For the structure of the received stabilizers, see the proof in [
16] (Page 926).
Lemma 3. Let s be a positive integer, and let Γ be a connected hexavalent -transitive graph for some . Let . Then and one of the following statements holds:
- (1)
For , the stabilizer is a -group.
- (2)
For , the stabilizer , , or .
- (3)
For , the stabilizer , , , . with and , or with and .
- (4)
For , the stabilizer .
3. A 2-arc Transitive Hexavalent Nonnormal Cayley Graph on
In this section, we construct a connected 2-arc transitive hexavalent nonnormal Cayley graph on and determine its full automorphism group. In fact, if is a Cayley graph of a non-abelian simple group G, then G is core free in X, where . Let and . Suppose that . Then by Lemma 3, n may be 60, 120, etc. Consider the action of X on the set of by right multiplication; then, . So, we may construct the nonnormal Cayley graph in , where , 120, etc. The following example is really the case where we construct .
Construction 1. Let G be the alternating group on the set . Then, . Let (the alternating group on ), where the following holds: Define .
Lemma 4. The graph in Construction 1 is a connected 2-arc-transitive graph and isomorphic to the nonnormal hexavalent Cayley graph of G, determined by with the following: Proof. Let
. Then,
X has a natural action on
. By Magma [
17],
, and so the graph
is connected by Lemma 2 (2). Furthermore, by Magma [
17], we have that
H is regular on
. However,
G is the stabilizer of point 1 in
X. Hence,
X has a factorization
with
. Therefore,
G is regular on
. By Lemma 2 (3),
is isomorphic to a Cayley graph of
. Additionally, by the computation of Magma [
17] (for the Magma code, see
Appendix A), we have
. Hence, Lemma 2 (1) implies that
is a hexavalent graph. Since
, Lemma 3 implies that
is 2-arc transitive. Since
X is a non-abelian simple group,
G is not normal in
. It follows that
is nonnormal. Let
,
,
,
,
and
S be defined as in this lemma. By the computation of Magma [
17] (for the Magma code, see
Appendix B), we have
. Thus, by Lemma 2 (3), we have that
is isomorphic to
. This completes the proof of the lemma. □
In the next lemma, we show that the full automorphism group is isomorphic to alternating group .
Lemma 5. The full automorphism group of the 2-arc-transitive hexavalent graph in Construction 1 is isomorphic to alternating group .
Proof. Let . Assume first that the full automorphism group is quasiprimitive on . Let N be a minimal normal subgroup of . Then, N is transitive on . It implies that N is insoluble. Thus, N is isomorphic to , where for each , T is a non-abelian simple group, and . Let p be the largest prime factor of the order of . Then, and . Since N is transitive on and , we have that p divides . Assume that . Then, divides . However, by Lemma 3, the order of the stabilizer divides , and so divides which is divisible by , a contradiction. Hence, we have and . Let be the centralizer of T in . Then, and . If , since is quasiprimitive on , this implies that C is transitive on . It implies that p divides . Therefore, divides , which divides , and so we have that divides , a contradiction. Hence, , and is almost simple.
Since
, it follows that
or
X. If
, then since
, we have
; note that
,
, a contradiction. Thus,
, and so
. It follows that
divides
, which divides
. By [
18] (pp. 135–136), we can conclude that
. Thus,
. If
, then
, a contradiction to Lemma 3. Hence,
.
Now assume that the full automorphism group is not quasiprimitive on . Then there is a minimal normal subgroup M of that acts nontransitively on . Since , we have or X. For the latter case , we have , and so M is transitive on , a contradiction. For the former case, , then we have that divides , which divides .
Assume that
M is insoluble. Since
divides
, and the simple groups
,
,
are the only
-factor non-abelian simple groups (see [
19] (Table 1), and note that the definition of the
-group is similar to
-group); by checking the orders of these groups, it is easy to figure out
or
or
. Then since
, we have
or
or
, a contradiction to the description of the orders of the stabilizers in Lemma 3.
Assume that
M is soluble. Then
or
or
, where
,
and
. Let
. Then
, a split expansion of
M by
X. Further, we have
or
or
. We note that
M is a subgroup of
. If
, then we have
or
or
. However, for each
,
and
,
,
or
has no subgroup isomorphic to the alternating group
. Hence, we have
and
. It implies that
; then
since
, we have
, and
X centralizes
M. Hence,
. Then
. Thus,
. Note that with the order of the stabilizers given in Lemma 3, we conclude
or
. In the case where
, we have
, then
,
, but there is no normal subgroup which is isomorphic to
in
, a contradiction. In the case where
, we have
, then
,
; by [
17], there is no normal subgroup with order 120 in
, so clearly,
, which also leads to a contradiction. This completes the proof of the lemma. □
Proof of Theorem 1. Now we are ready to prove our main Theorem 1. Let be the graph as in Construction 1. Then, Lemma 4 shows that is a connected 2-arc-transitive graph and isomorphic to a nonnormal hexavalent Cayley graph , with . This proves the statement of the former part of Theorem 1. The next Lemma 5 shows that the full automorphism group of the graph is isomorphic to alternating group . This proves the statement of the latter part of Theorem 1, and so completes the proof of Theorem 1. □