Abstract
This paper introduces the forbidden subgraph conditions for Hamilton-connected graphs. If the degree sequence of the graph is and it is connected, then it is called hourglass . For integers , the graph is paw, which is obtained by attaching one of the vertices of the triangle to one of the end vertices of a path with a number of edges i. We show that every graph G is Hamilton-connected if G is a -free, -free, -free, and a 3-connected graph. Moreover, we give an example to show the sharpness of a paw-type forbidden subgraph in a 3-connected, Hamilton-connected graph. Our focus on the Hamilton-connected problem can be applied to data center networks (DCNs). In the future, we will remove the forbidden subgraph families from our conclusions when building the network to obtain the optimal communication cost. Our result extends the result of Ryjáček and Vrána (Discrete Mathematics 344: 112350, 2021).
MSC:
05C40; 05C45; 05C75
1. Introduction
In this paper, for some undefined concepts, please refer to [1].The next part introduces common graph theory symbols and terminology. A graph always refers to a finite undirected simple graph, and is denoted by ; if multiedges (where at least two edges in an induced subgraph have the same pair of end vertices) are allowed, then the graph G is called a multigraph. If a graph G does not have an induced subgraph that is the same as the sets of vertices and edges of F, then it is called an F-free graph. is a cycle with a number of edges and edges i, and is a path with a number of vertices i. We use to denote the neighborhood of the vertex v in G, and the degree of a vertex v denotes the number of elements of . We denote the neighborhood and the degree of a vertex v in G with and , respectively. For simplicity, we can remove the subscript with G in this paper, e.g., we will write as . We use and to represent the maximum and the minimum degrees of a graph G, respectively.
Given a subgraph H of G, if the vertex set of H is the vertex set of G, then H is a spanning subgraph of G. Given the definition of a spanning subgraph, if G has a spanning cycle, then G is also Hamiltonian; for any pair vertices of G, there exists a spanning path with them as end vertices, and the graph G is Hamilton-connected. It is a difficult, well-studied, and meaningful problem to determine whether a graph is Hamilton-connected. They are central to graph theory. Some forbidden subgraphs used in this paper are shown in Figure 1. Theorem 1 lists known results on the Hamilton-connectedness of the forbidden subgraph pairs of the 3-connected graph.
Figure 1.
Some forbidden graphs.
Theorem 1.
Let G be -free and a 3-connected graph. Then:
- (1)
- (Ryjáček and Vrána [2]) G is Hamilton-connected if G is a -free graph of order .
- (2)
- (Bian et al. [3]) G is Hamilton-connected if G is a -free graph.
- (3)
- (Liu et al. [4]) G is Hamilton-connected if G is an -free graph.
We are very interested in the problem of forbidden subgraphs and, recently, many scholars have begun to pay attention to the problem of forbidden subgraph triples. Liu and Xiong [5] continued in this direction to prove Theorem 1 (2).
Theorem 2.
(Liu and Xiong [5]) Every -free, -free, -free, and 3-connected graph G is Hamilton-connected.
Liu and Xiong add a “-free” assumption condition in Theorem 2, which allows the results to be substantially strengthened. In this paper, according to the research direction of several researchers, we obtain a strengthened version of Theorem 1 (1), which adds the condition of “-free”.
Theorem 3.
Let a graph G be a -free, -free, -free, and 3-connected graph; then, it is Hamilton-connected.
In reality, the Hamilton-connected problem can be applied to communication networks where we remove the family of forbidden subgraphs in our conclusions when building the data center network; this is performed to better optimize the network cloud era and data center interconnection. For the application of other graphs, refer to [6].
In the study of any problem, an effective tool is always indispensable, and closure is a very effective mathematical model for dealing with the problem of the -free graph. In Section 2, we introduced tools and some important results for solving the forbidden subgraph problem, and the claw-free graph problem is no exception. Closure is a very effective mathematical model for dealing with forbidden subgraphs, especially claw-free graphs. The proof of Theorem 3 starts with some special edges; we consider their structure in the original graph, use their particularity, classify them, and prove the feasibility of the result one by one. It is deferred to Section 3.
2. Preliminaries
To make it easier to prove our result, we further introduce the following concepts.
Let , with S as the vertex set of the subgraph and all edges whose end vertices are in S, as the edge set is called the subgraph induced from the vertex set S (or simply induced subgraph), denoted by .
2.1. A Line Graph of Multigraph and Its Preimage
A line graph of G is a simple graph; its vertices are all edges of G and the edges e and f of G have a common vertex in G, which is the edge of the line graph, denoted as . If , then the preimage of H is G and .
If is a connected graph and is not complete, then the vertex x is eligible; all eligible vertices of G are placed in the set . And if is a complete graph, then x is simplicial. For given a vertex v in G, the graph resulting from adding all edges between two nonadjacent vertices in is called local completion . Clearly, when G is -free, so is .
When H is the preimage of G, The graph obtained by contracting the edge into a vertex and replacing the created loop with a new pendant edge is a line graph. Some properties of an eligible vertex are introduced in the following.
Lemma 1
(Ryjáček et al. [7]). If G is -free, where the center vertex of every induced of G is an eligible vertex, and , then the center vertex of every induced of is an eligible vertex.
Theorem 4
(Brousek and Ryjáček [8,9]). For and some vertex , if G is F-free and is also a -free graph, then so is .
2.2. SM-Closure
For a given -free graph G, the graph that can be obtained by recursive local completion of all eligible vertices of G is called its closure, for as long as possible. If , then G is closed. The graph constructed on -free graphs is introduced in [10].
- 1.
- If G is not Hamilton-connected, then all eligible vertices are locally completed in turn until the operation cannot proceed or the final graph that can be obtained is Hamilton-connected. This operation produces a series of graphs satisfying:
- (1)
- G is the starting graph ;
- (2)
- , where , for some eligible vertex of ;
- (3)
- For some vertices of , does not contain Hamiltonian -path; and
- (4)
- is Hamilton-connected with any eligible vertex x of , and set .
- 2.
- If G is Hamilton-connected, then we define as its closure.
The final resulting graph , obtained after the above multi-step iterative process, is called the -closure of G. Then, there is a definition of -closure that if a graph G is an -closure, then its equivalent is to say that this graph is -closed.
An -path is a path with as its end vertices. And -trail is a trail with as its terminal edges. The interior vertices in a trail T are denoted by the set Int. If an edge e has at least one vertex in vertex set , then Mdominatese. If all the edges of G are dominated by the trail T, then the closed trail T is called a dominated closed trail (DTC). If all the edges of G are dominated by Int, then -trail is called an internally dominating -trail (-IDT).
The graph G cannot uniquely determine its -closure. We will use some necessary features in the -closure as follows.
Theorem 5
(Kužel et al. [10]). Let be one of the -closures of the -free graph G. Then,
- (1)
- and ;
- (2)
- The fact that G is Hamilton-connected is equivalent to the fact that is Hamilton-connected;
- (3)
- The graph obtained by a series of local completions of G at eligible vertices is called ;
- (4)
- Given a graph H, is its line graph, that is, , satisfying:
- (A)
- H is simple graph and is also triangle-free, or;
- (B)
- For some edges , or H such that H is triangle-free and two edges are the only multiedges in H, there does not exist an -IDT, or at least one edge of is in a triangle of H, and H is a simple graph and has at most two triangles, and the triangles have no common edge if H has two triangles.
- (5)
- For some vertices of , let have no Hamiltonian path and -path. Then,
- (A)
- if is a multiedge.
- (B)
- if is a triangle.
Lemma 2
(Ryjáček and Vrána [11]). If G is a line graph of H and an -closed graph, then the triangle in H has no vertices with vertex degree 2.
Lemma 3
(Ryjáček et al. [7]). If be the -closure of a -free graph G, and let be a line graph of H, then v is an eligible vertex in and is equivalent to and is a multiedge or a triangle in H.
In the above lemmas, we introduce some results on the -closed graphs.
2.3. Strongly Spanning Trailable Multigraphs
The edges in H are replaced by paths , , respectively, where denote vertices; this type of graph is called multigraph . If any pair of edges (possible ) has a spanning -trail in multigraph H, then H is strongly spanning trailable (abbreviated SST).
Here, Wagner graph and are shown in Figure 2, where edges are multiedges. For example, in [12], Wagner graphs are known in the theory of networks-on-chip under the name of Spidergon topology and even implemented as an STNoC software solution. The line graphs of [13,14] strongly-spanning trailable graphs are known to be Hamilton-connected. In the graph G, the edge connectivity and circumference of G are denoted by the symbols and , respectively. In the following proof of Theorem 3, we also need to know the following results about SST.
Figure 2.
The graphs and .
Theorem 6
(Liu et al. [4]).
- (1)
- Let H be a 3-edge-connected multigraph of an order not greater than 9. Then, H is SST, except for the members of .
- (2)
- Let H be a 2-connected, 3-edge-connected multigraph where its circumference is at most 8. Then, H is SST, except for the members of .
If removing a(n) vertex(edge)-cut X from G results in at least two nontrivial components of G, then X is called essential; if there does not exist an essential vertex(edge)-cut X (the number of elements in X is less than k) in multigraph G, then G is called essentially k(-edge)-connected.
Given a multigraph H that is essentially 3-edge-connected, the resulting graph from the multigraph H replacing the path for each vertex y of degree 2 with an edge and deleting all the pendant vertices is called the core of H. In multigraph H, denotes the preimage of any subgraph F of , referring to the maximal subgraph of H obtained under the core operation. Next, we list some well-known results.
Theorem 7
(Shao, [15]). If H is a multigraph that is essentially 3-edge-connected, then the core of H has the following results:
- (1)
- is uniquely defined and ;
- (2)
- All edges of H are dominated by ;
- (3)
- is Hamilton-connected if is SST;
- (4)
- H has a DCT if has a spanning-closed trail.
3. Proof of Theorem 3
Let and be subsets of G such that ; if G is -free, then the further G is also -free. The shortest path in G connecting two vertices and is defined by , where , and we assume and to be oriented from to . If a subpath with the smaller number of vertices among the two subpaths with end vertices and in cycle C has l internal vertices and , then is a l-chord. An equivalent condition for a graph F is a subgraph of H is that the line graph of F is an induced subgraph of a line graph of G.
We use (for simply or ) in the resulting graph obtained by , subdividing its one edge times, where the labels of these vertices are given in Figure 3, and the vertex o will be called the center vertex.
Figure 3.
The graphs and .
Proof of Theorem 3.
Given a graph H, and letting G be its line graph, we suppose G is 3-connected and is not Hamilton-connected and let be the core of H. By Theorems 4 and 5 (3) and Lemma 1, we suppose that G is -closed, the center vertex of every induced in G is an eligible vertex. Suppose that the condition “G is a -free, -free and -free graph” of Theorem 3 is true. It suffices to show that H must contain a subgraph . By Theorem 7 (4), is not SST. By Theorem 7 (1),
Set
By Theorem 5 (4), either H is triangle-free and contains a multiedge, or H has no more than two triangles, and is a simple graph. Then, if H contains no more than two triangles, then and if H has a pair of multiedges, then . By Lemma 3, for an edge e in , the vertex does not correspond to the eligible vertex in G, there will exist e, which is not the center edge of a subgraph . Thus, we can get that
Claim 1.
and .
Proof.
Assume, to the contrary, that or . We know for a fact that the equivalent condition for every graph to be essentially 3-edge connected is that its line graph is 3-connected. By Theorem 6, . Then has a 8-cycle or 9-cycle with , and is a multiedge if there exists (see Figure 2).
By (1), for some integers , subdividing each edge of through a given vertex of degree 2, we can easily get that H contains subgraphs,
a contradiction. □
By Claim 1, we have that and . Next, using the following notation:
- always denotes a longest cycle of ;
- ;
- ;
- ;
- Let be the set of all edges between and . Then, .
Claim 2.
For , let be an -path, where , and let be a shorter -path with their vertices in . Then, (see Figure 4).
Figure 4.
The graph of Claim 2.
Proof.
Suppose Claim 2 is false, for some satisfying the hypothesis Claim 2. Then, is a cycle with no less than vertices, which contradicts the choice of . □
By (1) and the choice of , the longest cycle C of H is obtained by subdividing edges of the cycle by vertices of degree 2, respectively. Then, . For integers , we always use vertex to subdivide edge .
Claim 3.
.
Proof.
Suppose by contradiction. By (1), . Since , . Moreover, in addition to vertices and , there is at least one vertex v that satisfies . Then, H contains a subgraph,
a contradiction. □
Claim 4.
H has no multiedges.
Proof.
Assume, to the contrary, that H contains multiegdes. By Theorem 5 (4) (B), H contains at most two multiedges and no other multiedge. Let be a pair of multiedges, with as their end-vertices.
Suppose first that . In this case, , so we may assume and . Then is an -IDT in H, contradicting Theorem 5 (5) (A). Now suppose that .
Firstly, suppose that , but then , contradicting .
Then, suppose that . By Claim 3, . We suppose that and . Since H is triangle-free, is a l-chord of with , we get that , a contradiction. Then, suppose that and is not in , by (1), . Then, H contains a subgraph,
a contradiction.
Finally, suppose that , say . By (1), . Moreover, there is at least one edge in with as its end-vertex should be a subdivided of degree 2 in H. Then, H contains a subgraph,
with its center vertex , a contradiction.
Thus H has no multiedges. □
By Claims 3 and 4, we obtain that H is a simple graph.
Claim 5.
H contains at least one triangle and is a simple graph.
Proof.
Assume that H is triangle-free and is a simple graph by contradiction. Since , the degree of is at least 3. By (1), . Then, H contains a subgraph,
a contradiction. □
By Claim 5, H contains either one triangle or two triangles. By Lemma 2, each vertex of the triangle in H has a degree of at least 3. The definition of shows that the edges of the triangle in H are also the edges of the triangle in .
Claim 6.
Let be a triangle of . Then,
- (1)
- with .
- (2)
- if contains at least one triangle.
Proof.
(1) Suppose that, for any , by contradiction. And since , and , then , a contradiction.
(2) We only need to prove that . Firstly, suppose that , but then , contradicting . Therefore, , and we can easily get that .
Then, suppose that . By Claim 3, . Then with , which contradicts , a contradiction. Therefore, .
Finally, suppose that , let . By Claim 2, . Then , a contradiction. Therefore, . □
Next, we consider the case of . For convenience, suppose that the vertices of a triangle in are denoted by and (see Figure 5). By Claim 6 (1), we have that . Then, . By (1), . Since , suppose that and such that . Then, H contains a subgraph,
a contradiction. Therefore, .
Figure 5.
A cycle C of circumference at least 9 in and .
Choose a shortest path in H with their vertices not in C and . Then, we consider the case of . In the following proof of Theorem 3, we denote two triangles by and , by symmetry, . By Claim 6 (1), we have that and .
Claim 7.
Suppose that and . Then, .
Proof.
Assume, to the contrary, that (say ). Then , and and . By (1), .
Case 1.
(see Figure 6(1)).
Figure 6.
The cycle C with a circumference of 9 in and .
Suppose that . Then, there exists a path in H for any possibility and H contains a subgraph
a contradiction. Hence, is an end vertex of the chord in , that is,
Suppose that . Then, there exists a path in H for any possibility and H contain a subgraph,
a contradiction. Hence, is an end vertex of the chord in (). Similarly, . Hence, by Claim 6 (1), and , contradicting (2). This proves Case 1.
Case 2.
(see Figure 6 (2)).
Suppose that . Then there exists a path in H for any possibility , and H contains a subgraph,
a contradiction. Hence, is an end vertex of the chord in . Similarly, is an end vertex of the chord in . Next, suppose that . Then there exists a path in H for any possibility , and H contains a subgraph,
a contradiction. Hence, is an end vertex of the chord in . Similarly, is an end vertex of the chord in . Then we have that is an end vertex of the chord in , Then, we have that lies on a chord of , otherwise H contains a subgraph , , a contradiction. Therefore, , a contradiction. □
By Claims 1 and 4–7, we can get that and and , where . By (1), . Since , . We can get that H contains a subgraph such that there are two edges with and . By (1), should be subdivided by two vertex of degree 2, say , respectively. We have that H contains a subgraph,
a contradiction. This proves Theorem 3. □
4. Conclusions
Remark 1.
As shown in Figure 7, we give a family of graphs . Then, any graph is a 3-connected, -free, and -free graph, and G is not Hamilton-connected. It is easy to see that the graphs in Figure 7 contain a subgraph , but they are -free. Thus, these examples of these graphs presented in Figure 7 show that our result of Theorem 3 “-free” is sharp.
Figure 7.
The family of graphs .
Remark 2.
In this paper, we are mainly concerned with the effect of an induced hourglass on existing results. We list the results to show how to effect a graph to be Hamilton-connected, when we add the -free condition to 3-connected, -free, and -free graphs. Our results show that the subscript 14 () is almost double the original one 7 ().
Remark 3.
In Figure 7, we can easily find that these graphs are also -free, -free, and -free with integers . Also, we guess Theorem 1 (3) that adds the condition “-free”, and the following problem may occur with the forbidden subgraphs condition of the Net-type.
Question 1.
Let G be a 3-connected graph. If G is a -free, -free, and -free graph, then G is Hamilton-connected.
5. Discussion
The Hamilton-connected problem has many applications in computer science, communication networks, and coding theory, and they can help create a hierarchical organization; this organization does not contain forbidden subgraphs in our results, which determines optimal communication cost, fastest computer response time, or the shortest distance of a computer network. This paper discussed the sufficient conditions for the existence of Hamilton-connection. By combining the paw-type stability with Hamilton-connected graphs, we aim to build a suitable mathematical model and apply it to reality, in order to lay a foundation and theoretical support.
Author Contributions
Conceptualization, P.W. and L.X.; Methodology, P.W. and L.X.; Writing—Original draft preparation, P.W.; Writing—Review and Editing, P.W. and L.X.; Funding acquisition, L.X. All authors have read and agreed to the published version of the manuscript.
Funding
This work was funded by National Natural Science Foundation of China of grant number 12131013.
Data Availability Statement
Data are contained within the article.
Acknowledgments
We are very grateful to the reviewers for their many revisions and a large number of useful suggestions, which have improved the quality of our article.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Bondy, J.A.; Murty, U.S.R. Graph Theory; Springer: Berlin/Heidelberg, Germany, 2008. [Google Scholar]
- Ryjáček, Z.; Vrána, P. Every 3-connected {K1,3,Z7}-free graph of order at least 21 is Hamilton-connected. Discret. Math. 2021, 344, 112350. [Google Scholar] [CrossRef]
- Bian, Q.; Gould, R.; Horn, P.; Janiszewski, S.; Fleur, S.; Wrayno, P. 3-connected {K1,3,P9}-free graphs are Hamiltonian-connected. Graphs Comb. 2014, 30, 1099–1122. [Google Scholar] [CrossRef]
- Liu, X.; Xiong, L.; Lai, H.-J. Strongly spanning trailable graphs with small circumference and Hamilton-connected claw-free graphs. Graphs Comb. 2021, 37, 65–85. [Google Scholar] [CrossRef]
- Liu, X.; Xiong, L. A note on 3-connected hourglsss-free claw-free Hamilton-connected graphs. Discret. Math. 2022, 345, 112910. [Google Scholar] [CrossRef]
- Lv, M.; Fan, J.; Fan, W.; Jia, X. A high-performantal and server-centric based data center network. IEEE Trans. Netw. Sci. Eng. 2023, 10, 592–605. [Google Scholar] [CrossRef]
- Ryjáček, Z.; Vrána, P.; Xiong, L. Hamiltonian properties of 3-connected {claw,hourglass}-free graphs. Discret. Math. 2018, 341, 1806–1815. [Google Scholar] [CrossRef]
- Brousek, J.; Ryjáček, Z.; Favaron, O. Forbidden subgraphs, hamiltonicity and closure in claw-free graphs. Discret. Math. 1999, 196, 29–50. [Google Scholar] [CrossRef]
- Brousek, J.; Ryjáček, Z.; Schiermeyer, I. Forbidden subgraphs, stability and hamiltonicity. Discret. Math. 1999, 197/198, 143–155. [Google Scholar] [CrossRef]
- Kužel, R.; Ryjáček, Z.; Teska, J.; Vrána, P. Closure, clique covering and degree conditions for Hamilton-connectedness in claw-free graphs. Discret. Math. 2012, 312, 2177–2189. [Google Scholar] [CrossRef]
- Ryjáček, Z.; Vrána, P. A closure for 1-Hamilton-connectedness in claw-free graphs. J. Graph Theory 2014, 75, 358–376. [Google Scholar] [CrossRef]
- Tatas, K.; Siozios, K.; Soudris, D.; Jantsch, A. Designing 2D and 3D Network-on-Chip Architectures; Springer: Berlin/Heidelberg, Germany, 2014. [Google Scholar]
- Lai, H.-J.; Shao, Y.; Yu, G.; Zhan, M. Hamiltonian connectedness in 3-connected line graphs. Discret. Math. 2009, 157, 152–173. [Google Scholar] [CrossRef]
- Lui, J.; Yu, A.; Wang, K.; Lai, H.-J. Degree sum and Hamiltonian-connected line graphs. Discret. Math. 2018, 341, 1363–1379. [Google Scholar]
- Shao, Y. Claw-Free Graphs and Line Graphs. Ph.D. Thesis, West Virginia University, Morgantown, WV, USA, 2005. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).






