1. Introduction
In the era of internet-of-things (IoT), wireless sensor networks (WSNs) are utilised in many applications, including smart grids (SGs) [
1]. When the number of sensors continues to increase, faulty nodes and faulty links are inevitable in wireless sensor networks, which undoubtedly increase the difficulty of network reliability assessment and will adversely affect its fault tolerance and reliability.
Hence, to the design and maintenance purpose of networks, appropriate measures of reliability should be found. Because the topology of network system can usually be represented by a connected graph
, where
and
can be represented as the set of processors and the set of communication links between processors, respectively. Connectivity
and edge-connectivity
represent the maximum number of faulty vertices and the maximum number of faulty edges that an interconnected network can tolerate in normal communication, respectively. Obviously, the higher the (edge-) connectivity is, the more reliable an interconnection network is [
2]. As is known,
, where
is the minimum degree of
G. Boesch and Tindell [
3] introduced the concepts of super connectivity and super edge- connectivity. A connected graph
G is super connected (resp., edge-connected, in short super-
(resp., super-
, ) if any minimum vertex-cut (resp., edge-cut) isolates a vertex of
G.
However, the biggest disadvantage of connectivity and edge-connectivity in reliability evaluation and fault-tolerant performance of networks is that they all default to a vertex where all adjacent vertices or adjacent edges fail at the same time, which rarely happens in the real world. To overcome such shortcomings, several new concepts on the (edge-) connectivity of graphs, which are called conditional (edge-) connectivity, were proposed by Harary [
4]. Among them are restricted (edge-)connectivity and cyclic edge-connectivity.
Esfahanian and Hakimi proposed the concept of restricted connectivity [
5]. For a given non-negative integer
h, a vertex-set (resp., edge-set)
F is an
h-restricted vertex-cut (resp., edge-cut) of a connected graph
G if
is disconnected, and each component of
contains at least
vertices. The minimum size of all
h-restricted vertex-cuts (resp., edge-cuts), denoted by
(resp.,
), is called
h-restricted connectivity (resp., edge-connectivity) of the graph
G. In the exceptional cases,
G is called super restricted vertex (edge)-connected and denoted by super-
(resp., super-
), if every minimum restricted edge cut isolates one component of order 2. The
h-restricted (edge)-connectivity and super-
(resp., super-
) graph for some classic interconnection networks have been investigated recently. For instance, Li et al. [
3] determine that the
h-restricted connectivity of the arrangement graph
is
for
h is odd and
, otherwise the
h-restricted connectivity of the arrangement graph
is
for
and
. Chen et al. [
6] proved that the hypercube, the twisted cube, the cross cube, the Möbius cube, and the locally twisted cube are super connected and super edge-connected. Futhermore, Xu et al. [
7] obtained super edge-connectivity, and the restricted (edge-) connectivity of the hypercube, the twisted cube, the cross cube, Möbius cube, and the locally twisted cube is
for
. Zhou et al. [
8] obtained that the balanced hypercube
is super-
but not super-
for
.
In 1978, Bollobás indicated all multigraphs without two vertex-disjoint cycles [
9]. So, it is necessary to further study the cyclically separable graphs. An edge set
S is a cyclic edge-cut if
is disconnected and at least two of its components contain cycles. Obviously, a graph
G has a cyclic edge-cut if and only if it has two vertex-disjoint cycles and is a cyclically separable graphs. For a cyclically separable graph
G, the cyclic edge-connectivity of
G, denoted by
, is defined as the cardinality of the minimum cyclic edge-cut of
G [
10]. Let
be the number of edges with one end in
X and another end in
, and
. A cyclically separable graph
G with
is called cyclically optimal. Moreover, if removing any minimum cyclic edge-cut of the graph
G results in the shortest cycle in a component of the graph
G, then the graph
G is said to be super cyclically edge-connected [
11], in short, super-
. There are some results of the super cyclically edge-connectivity for some classes of the graphs. For example, Zhang [
12] showed that the cyclic edge-connectivity of a strongly regular connected graph
G (not
) of degree
with girth
g is equal to
, where
. Wang and Zhang [
13] showed that any
k-regular vertex-transitive connected graph with
and girth of at least 5 is cyclically optimal, and any edge-transitive connected graph with minimum degree of at least 4 and order of at least 6 is cyclically optimal.
Hypercube
, one of the most important and attractive network topologies so far, is widely used due to its excellent characteristics. In 1994, Cheng and Chuang designed a network topology capable of obtaining an interconnection network that is both efficient and cost-effective, this is, the varietal hypercube
. In recent years,
has received considerable attention and many of its properties have been studied. For example, Zhou [
14] showed that
is a Cayley graph. Wang et al. [
15] also determined the full automorphism group of the varietal hypercube
. More relevant results on the varietal hypercube
are shown in [
16,
17,
18,
19,
20].
Our work in this paper concerns the n-dimensional varietal hypercube . We use another method to prove that is super vertex-connected for any and super edge-connected for . We also obtain that is super- for , super- for and , and the cyclic edge-connectivity of is for .
The paper is organized as follows. In
Section 2, we give some propositions related to the vertex-transitive and edge-transitive graph. In
Section 3, we introduce the necessary definitions and notations. In
Section 4, we use the symmetry properties of the
n-dimensional varietal hypercube to discuss some reliability measures of the
n-dimensional varietal hypercube. In
Section 5, we add some concluding remarks.
3. Preliminaries
In this section, we give some definitions and notations. If definitions and notations are not defined here, we follow [
29].
Let be the automorphism group of the graph G. We say that G is vertex-transitive if for any two vertices there exists an automorphism such that . If for any pair of edges (resp. arcs), there exists an automorphism that transforms one into the other, then G is edge-transitive (resp. arc-transitive) graph. Clearly, if a graph is arc-transitive, then it must be vertex-transitive and edge-transitive. Let u and v be two vertices of G, and the length of a shortest path is called the distance and denoted by . The maximum distance between any pair of vertices of a connected graph G is called the diameter of the graph G and denoted by .
Suppose that
G and
H are two graphs, the lexicographic product
is defined as the graph with vertex set
, two vertices
and
are adjacent if and only if either
, or
and
. The graph showed in
Figure 1 is
.
Let n be a positive integer. The definitions of the n-dimensional hypercube and the n-dimensional varietal hypercube are stated as follows.
Definition 1 ([
30])
. The n-dimensional hypercube is a connected graph with vertices and denoted by . The vertex set . Two vertices and in are adjacent if and only if they differ in exact one position. Definition 2 ([
17])
. The n-dimensional varietal hypercube, denoted by , has vertices, each labeled by an n-bit binary string and . is a complete graph of two vertices labeled with 0 and 1, respectively. For , can be recursively constructed from two copies of , denoted by and , and by adding edges between and , where , . The vertex is adjacent to the vertex if and only ifObviously, is an n-regular and its girth is 4. Moreover, it contains circles of length 5 when [18,19]. The varietal hypercubes , , and are illustrated in Figure 2. As a variant of the hypercube, the
n-dimensional varietal hypercubes
, which has the same number of vertices and edges as
, not only has the most ideal characteristics of
, including some characteristics such as recursive structure, strong connectivity, and symmetry but also has a smaller diameter than
, and its average distance is smaller than the hypercube [
17].
Lemma 1 ([
17])
. for . Lemma 2 ([
14])
. for . Lemma 3 ([
15,
20,
31])
. is vertex-transitive and edge-transitive. Lemma 4 ([
32])
. Let G and H be two graphs. If G is non-trivial, non-complete, and connected, then . Lemma 5 ([
33]).
Suppose and are two vertices of . Then, the distance between and is From the above lemma, we can obtain that the diameter of
is
4. Reliability Evaluation of Varietal Hypercubes
In this section, we will study some reliability measures in varietal hypercubes, such as restricted (vertex-)connectivity and cyclic connectivity. Although Xu et al. proved that
is super vertex-connected for any
and super edge-connected for
in [
19], we prove these results and obtain other results such as
is super-
for
, super-
for
, and
through symmetric properties of
.
Theorem 1. is super-κ for .
Proof. If , then is a complete graph . Obviously, it is super-.
If , then is cycle , and clearly it is super-.
Assume that . Suppose to the contrary that is not super-. Since is both vertex-transitive and edge-transitive, it follows that from Proposition 1.
First suppose that ; we know that and by Lemma 4, and . Moreover, since , it follows that . Thus, m and l are even. According to Lemma 1 and Lemma 5, we know that and , respectively. Now, and . Now, , and we have .
We discuss the value of m, as follows.
Case 1. where k is a positive integer.
Then, . Now, we obtain , a contradiction.
Case 2. where k is a positive integer.
Then, . So, we have . In this case, ; hence, we have and . This contradicts the fact that m is even.
Case 3. where k is a positive integer.
Then, . So, we have and , a contradiction.
Now, suppose that , then the number of vertices of these graphs is same and we obtain , so we have and where , a contradiction. □
Theorem 2. is super-λ for .
Proof. Since is a complete graph , it is super-. If , is a cycle . Clearly, it is not super-. Suppose that is not super- for . is edge-transitive k-regular graph where , so the theorem follows from Proposition 2(1). □
In [
19], Xu et al. proved that 1-restricted edge-connectivity of
is
. Here, we present another proof of this result.
Theorem 3. for .
Proof. If , then is a cycle . Clearly, . Thus, we may suppose that . is an edge-transitive k-regular graph where , so the theorem follows from Proposition 2(2). □
In the following theorem, we show that every minimum edge-cut of for and isolates an edge.
Theorem 4. is super- for and .
Proof. Assume that is not super- for and . According to is edge-transitive and Proposition 2(4), we know that is not super- if and only if is isomorphic to or to an edge-transitive 4-regular graph of girth 4.
If , does not have cycle of length 5 but has, then is not isomorphic to .
If , since is edge-transitive k-regular graph where , then it implies that is super- when .
We have accordingly proved the theorem. □
In the following theorem, we prove that if we remove edges from the , then we will obtain a disconnected graph with at least two components containing cycle for .
Theorem 5. for .
Proof. Since has at least 8 vertices for , it is not isomorphic to , or . According to the fact that has girth 4, it follows that by Proposition 2(3). □
The following theorem shows that every minimum cyclic edge-cut of isolates a shortest cycle for .
Theorem 6. is super- for .
Proof. Since is an edge-transitive n-regular graph of girth with vertices. By the definition of , we obtain that is not isomorphic to , or . If , then it is obvious that is super-.
It is assumed that is not super- for . So if , then is edge-transitive 4-regular graph. According to Proposition 3(3), it follows that is isomorphic to (the 4-cube), , , or , for some arc-transitive 4-regular graph H. Since has odd cycle of length 5 and 16 vertices, it is not difficult to check that is not isomorphic to , , or , for some 4-regular arc-transitive graph H. Furthermore, if for and , then , it follows that and . However, in this case, and by Lemma 1 and Lemma 5. So, is not isomorphic to , a contradiction. Now, by Proposition 3(3) and (4), or has 12 or 16 vertices, a contradiction.
Thus, we conclude that is super- for . □
The next theorem shows that every minimum vertex-cut of isolates an edge for .
Theorem 7. is super- for .
Proof. By Lemma 2, we know that for . It is assumed that is not super- for . Hence, there is an 1-restricted vertex-cut S and , but the cut is not the neighborhood of any edge. That is, S is also a two-restricted vertex-cut and . So, we have , and this leads to a contradiction. □