arXiv subjects

Songling Shan

Publications and source records attributed to Songling Shan.

72 records · Page 4Linked to original sources

arXiv2019

Nonempty intersection of longest paths in graphs without forbidden pairs

In 1966, Gallai asked whether all longest paths in a connected graph have a nonempty intersection. The answer to this question is not true in general and various counterexamples have been found. However, there is a positive solution to Gallai's question for many well-known classes of graphs such as split graphs, series parallel graphs, and $2K_2$-free graphs. Among all the graph classes that support Gallai's question, almost all of them were shown to be Hamiltonian under certain conditions. This observation motivates us to investigate Gallai's question in graphs that are "close" to Hamiltonicity properties. Let $\{R,S\}$ be a pair of connected graphs. In particular, in this paper, we show that Gallai's question is affirmative for all connected $\{R,S\}$-free graphs such that every 2-connected $\{R,S\}$-free graph is Hamiltonian. These pairs $\{R,S\}$ were completely characterized in 1990s.

math.CO↗

arXiv2019

Toughness and spanning trees in $K_4$-minor-free graphs

For an integer $k$, a $k$-tree is a tree with maximum degree at most $k$. More generally, if $f$ is an integer-valued function on vertices, an $f$-tree is a tree in which each vertex $v$ has degree at most $f(v)$. Let $c(G)$ denote the number of components of a graph $G$. We show that if $G$ is a connected $K_4$-minor-free graph and $$ c(G-S) \;\le\; \sum_{v \in S} (f(v)-1) \quad\hbox{for all $S \subseteq V(G)$ with $S \ne \emptyset$} $$ then $G$ has a spanning $f$-tree. Consequently, if $G$ is a $\frac{1}{k-1}$-tough $K_4$-minor-free graph, then $G$ has a spanning $k$-tree. These results are stronger than results for general graphs due to Win (for $k$-trees) and Ellingham, Nam and Voss (for $f$-trees). The $K_4$-minor-free graphs form a subclass of planar graphs, and are identical to graphs of treewidth at most $2$, and also to graphs whose blocks are series-parallel. We provide examples to show that the inequality above cannot be relaxed by adding $1$ to the right-hand side, and also to show that our result does not hold for general planar graphs. Our proof uses a technique where we incorporate toughness-related information into weights associated with vertices and cutsets.

math.CO↗

arXiv2019

Hamiltonian cycles in tough $(P_2\cup P_3)$-free graphs

Let $t>0$ be a real number and $G$ be a graph. We say $G$ is $t$-tough if for every cutset $S$ of $G$, the ratio of $|S|$ to the number of components of $G-S$ is at least $t$. Determining toughness is an NP-hard problem for arbitrary graphs. The Toughness Conjecture of Chvátal, stating that there exists a constant $t_0$ such that every $t_0$-tough graph with at least three vertices is hamiltonian, is still open in general. A graph is called $(P_2\cup P_3)$-free if it does not contain any induced subgraph isomorphic to $P_2\cup P_3$, the union of two vertex-disjoint paths of order 2 and 3, respectively. In this paper, we show that every 15-tough $(P_2\cup P_3)$-free graph with at least three vertices is hamiltonian.

math.CO↗

arXiv2019

Toughness and prism-hamiltonicity of $P_4$-free graphs

The \emph{prism} over a graph $G$ is the product $G \Box K_2$, i.e., the graph obtained by taking two copies of $G$ and adding a perfect matching joining the two copies of each vertex by an edge. The graph $G$ is called \emph{prism-hamiltonian} if it has a hamiltonian prism. Jung showed that every $1$-tough $P_4$-free graph with at least three vertices is hamiltonian. In this paper, we extend this to observe that for $k \geq 1$ a $P_4$-free graph has a spanning \emph{$k$-walk} (closed walk using each vertex at most $k$ times) if and only if it is $\frac{1}{k}$-tough. As our main result, we show that for the class of $P_4$-free graphs, the three properties of being prism-hamiltonian, having a spanning $2$-walk, and being $\frac{1}{2}$-tough are all equivalent.

math.CO↗

arXiv2018

Independence number of edge-chromatic critical graphs

Let $G$ be a simple graph with maximum degree $Δ(G)$ and chromatic index $χ'(G)$. A classic result of Vizing indicates that either $χ'(G )=Δ(G)$ or $χ'(G )=Δ(G)+1$. The graph $G$ is called $Δ$-critical if $G$ is connected, $χ'(G )=Δ(G)+1$ and for any $e\in E(G)$, $χ'(G-e)=Δ(G)$. Let $G$ be an $n$-vertex $Δ$-critical graph. Vizing conjectured that $α(G)$, the independence number of $G$, is at most $\frac{n}{2}$. The current best result on this conjecture, shown by Woodall, is that $α(G)<\frac{3n}{5}$. We show that for any given $\varepsilon\in (0,1)$, there exist positive constants $d_0(\varepsilon)$ and $D_0(\varepsilon)$ such that if $G$ is an $n$-vertex $Δ$-critical graph with minimum degree at least $d_0$ and maximum degree at least $D_0$, then $α(G)<(\frac{1}{2}+\varepsilon)n$. In particular, we show that if $G$ is an $n$-vertex $Δ$-critical graph with minimum degree at least $d$ and $Δ(G)\ge (d+2)^{5d+10}$, then \[ α(G) < \left. \begin{cases} \frac{7n}{12}, & \text{if $d= 3$; } \frac{4n}{7}, & \text{if $d= 4$; } \frac{d+2+\sqrt[3]{(d-1)d}}{2d+4+\sqrt[3]{(d-1)d}}n<\frac{4n}{7}, & \text{if $d\ge 19$. } \end{cases} \right. \]

math.CO↗

arXiv2017

Vizing's 2-factor Conjecture Involving Toughness and Maximum Degree Conditions

Let $G$ be a simple graph, and let $Δ(G)$ and $χ'(G)$ denote the maximum degree and chromatic index of $G$, respectively. Vizing proved that $χ'(G)=Δ(G)$ or $Δ(G)+1$. We say $G$ is $Δ$-critical if $χ'(G)=Δ+1$ and $χ'(H)<χ'(G)$ for every proper subgraph $H$ of $G$. In 1968, Vizing conjectured that if $G$ is a $Δ$-critical graph, then $G$ has a 2-factor. Let $G$ be an $n$-vertex $Δ$-critical graph. It was proved that if $Δ(G)\ge n/2$, then $G$ has a 2-factor; and that if $Δ(G)\ge 2n/3+12$, then $G$ has a hamiltonian cycle, and thus a 2-factor. It is well known that every 2-tough graph with at least three vertices has a 2-factor. We investigate the existence of a 2-factor in a $Δ$-critical graph under "moderate" given toughness and maximum degree conditions. In particular, we show that if $G$ is an $n$-vertex $Δ$-critical graph with toughness at least 3/2 and with maximum degree at least $n/3$, then $G$ has a 2-factor. In addition, we develop new techniques in proving the existence of 2-factors in graphs.

math.CO↗

arXiv2017

Dirac's Condition for Spanning Halin Subgraphs

Let $G$ be an $n$-vertex graph with $n\ge 3$. A classic result of Dirac from 1952 asserts that $G$ is hamiltonian if $δ(G)\ge n/2$. Dirac's theorem is one of the most influential results in the study of hamiltonicity and by now there are many related known results\,(see, e.g., J. A. Bondy, Basic Graph Theory: Paths and Circuits, Chapter 1 in: {\it Handbook of Combinatorics Vol.1}). A {\it Halin graph} is a planar graph consisting of two edge-disjoint subgraphs: a spanning tree of at least 4 vertices and with no vertex of degree 2, and a cycle induced on the set of the leaves of the spanning tree. Halin graphs possess rich hamiltonicity properties such as being hamiltonian, hamiltonian connected, and almost pancyclic. As a continuous "generalization" of Dirac's theorem, in this paper, we show that there exists a positive integer $n_0$ such that any graph $G$ with $n\ge n_0$ vertices and $δ(G)\ge (n+1)/2$ contains a spanning Halin subgraph. In particular, it contains a spanning Halin subgraph which is also pancyclic.

math.CO↗

arXiv2017

Hamiltonian cycles in 3-tough $2K_2$-free graphs

A graph is called $2K_2$-free if it does not contain two independent edges as an induced subgraph. Broersma, Patel, and Pyatkin showed that every 25-tough $2K_2$-free graph with at least three vertices is hamiltonian. In this paper, we improve the required toughness in this result from 25 to 3.

math.CO↗

arXiv2017

Antimagic orientation of biregular bipartite graphs

An antimagic labeling of a directed graph $D$ with $n$ vertices and $m$ arcs is a bijection from the set of arcs of $D$ to the integers $\{1, \cdots, m\}$ such that all $n$ oriented vertex sums are pairwise distinct, where an oriented vertex sum is the sum of labels of all arcs entering that vertex minus the sum of labels of all arcs leaving it. An undirected graph $G$ is said to have an antimagic orientation if $G$ has an orientation which admits an antimagic labeling. Hefetz, M{ü}tze, and Schwartz conjectured that every connected undirected graph admits an antimagic orientation. In this paper, we support this conjecture by proving that every biregular bipartite graph admits an antimagic orientation.

math.CO↗

arXiv2016

Chromatic index determined by fractional chromatic index

Given a graph $G$ possibly with multiple edges but no loops, denote by $Δ$ the {\it maximum degree}, $μ$ the {\it multiplicity}, $χ'$ the {\it chromatic index} and $χ_f'$ the {\it fractional chromatic index} of $G$, respectively. It is known that $Δ\le χ_f' \le χ' \le Δ+ μ$, where the upper bound is a classic result of Vizing. While deciding the exact value of $χ'$ is a classic NP-complete problem, the computing of $χ_f'$ is in polynomial time. In fact, it is shown that if $χ_f' > Δ$ then $χ_f'= \max \frac{|E(H)|}{\lfloor |V(H)|/2\rfloor}$, where the maximality is over all induced subgraphs $H$ of $G$. Gupta\,(1967), Goldberg\,(1973), Andersen\,(1977), and Seymour\,(1979) conjectured that $χ'=\lceilχ_f'\rceil$ if $χ'\ge Δ+2$, which is commonly referred as Goldberg's conjecture. In this paper, we show that if $χ' >Δ+\sqrt[3]{Δ/2}$ then $χ'=\lceilχ_f'\rceil$. The previous best known result is for graphs with $χ'> Δ+\sqrt{Δ/2}$ obtained by Scheide, and by Chen, Yu and Zang, independently. It has been shown that Goldberg's conjecture is equivalent to the following conjecture of Jakobsen: {\it For any positive integer $m$ with $m\ge 3$, every graph $G$ with $χ'>\frac{m}{m-1}Δ+\frac{m-3}{m-1}$ satisfies $χ'=\lceilχ_f'\rceil$.} Jakobsen's conjecture has been verified for $m$ up to 15 by various researchers in the last four decades. We show that it is true for $m\le 23$. Moreover, we show that Goldberg's conjecture holds for graphs $G$ with $Δ\leq 23$ or $|V(G)|\leq 23$.

math.CO↗

arXiv2016

Nonempty intersection of longest paths in $2K_2$-free graphs

In 1966, Gallai asked whether all longest paths in a connected graph share a common vertex. Counterexamples indicate that this is not true in general. However, Gallai's question is positive for certain well-known classes of connected graphs, such as split graphs, interval graphs, circular arc graphs, outerplanar graphs, and series-parallel graphs. A graph is $2K_2$-free if it does not contain two independent edges as an induced subgraph. In this paper, we show that in nonempty $2K_2$-free graphs, every vertex of maximum degree is common to all longest paths. Our result implies that all longest paths in a nonempty $2K_2$-free graph have a nonempty intersection. In particular, it gives a new proof for the result on split graphs, as split graphs are $2K_2$-free.

math.CO↗

arXiv2016

Spanning trails with maximum degree at most 4 in $2K_2$-free graphs

A graph is called $2K_2$-free if it does not contain two independent edges as an induced subgraph. Mou and Pasechnik conjectured that every $\frac{3}{2}$-tough $2K_2$-free graph with at least three vertices has a spanning trail with maximum degree at most $4$. In this paper, we confirm this conjecture. We also provide examples for all $t < \frac{5}{4}$ of $t$-tough graphs that do not have a spanning trail with maximum degree at most $4$.

math.CO↗

arXiv2016

Regular colorings and factors of regular graphs

An $(r-1,1)$-coloring of an $r$-regular graph $G$ is an edge coloring such that each vertex is incident to $r-1$ edges of one color and $1$ edge of a different color. In this paper, we completely characterize all $4$-regular pseudographs (graphs that may contain parallel edges and loops) which do not have a $(3,1)$-coloring. An $\{r-1,1\}$-factor of an $r$-regular graph is a spanning subgraph in which each vertex has degree either $r-1$ or $1$. We prove various conditions that that must hold for any vertex-minimal $5$-regular pseudographs without $(4,1)$-colorings or without $\{4,1\}$-factors. Finally, for each $r\geq 6$ we construct graphs that are not $(r-1,1)$-colorable and, more generally, are not $(r-t,t)$-colorable for small $t$.

math.CO↗

arXiv2016

A Note On Vertex Distinguishing Edge colorings of Trees

A proper edge coloring of a simple graph $G$ is called a vertex distinguishing edge coloring (vdec) if for any two distinct vertices $u$ and $v$ of $G$, the set of the colors assigned to the edges incident to $u$ differs from the set of the colors assigned to the edges incident to $v$. The minimum number of colors required for all vdecs of $G$ is denoted by $χ\,'_s(G)$ called the vdec chromatic number of $G$. Let $n_d(G)$ denote the number of vertices of degree $d$ in $G$. In this note, we show that a tree $T$ with $n_2(T)\leq n_1(T)$ holds $χ\,'_s(T)=n_1(T)+1$ if its diameter $D(T)=3$ or one of two particular trees with $D(T) =4$, and $χ\,'_s(T)=n_1(T)$ otherwise; furthermore $χ\,'_{es}(T)=χ\,'_s(T)$ when $|E(T)|\leq 2(n_1(T)+1)$, where $χ\,'_{es}(T)$ is the equitable vdec chromatic number of $T$.

math.CO↗

arXiv2015

Difference of forbidden pairs containing a claw

When we study forbidden subgraph conditions guaranteeing graphs to have some properties, a claw (or $K_{1,3}$) frequently appears as one of forbidden subgraphs. Recently, Furuya and Tsuchiya compared two classes generated by different forbidden pairs containing a claw, and characterized one of such classes. In this paper, we give such characterization for three new classes. Furthermore, we give applications of our characterizations to some forbidden subgraph problems.

math.CO↗

arXiv2014

Minimum degree condition for spanning generalized Halin graphs

A spanning tree with no vertices of degree 2 is called a Homeomorphically irreducible spanning tree\,(HIST). Based on a HIST embedded in the plane, a Halin graph is formed by connecting the leaves of the tree into a cycle following the cyclic order determined by the embedding. Both of the determination problems of whether a graph contains a HIST or whether a graph contains a spanning Halin graph are shown to be NP-complete. It was conjectured by Albertson, Berman, Hutchinson, and Thomassen in 1990 that a {\it every surface triangulation of at least four vertices contains a HIST}\,(confirmed). And it was conjectured by Lovász and Plummer that {\it every 4-connected plane triangulation contains a spanning Halin graph}\,(disproved). Balancing the above two facts, in this paper, we consider generalized Halin graphs, a family of graph structures which are "stronger" than HISTs but "weaker" than Halin graphs in the sense of their construction constraints. To be exact, a generalized Halin graph is formed from a HIST by connecting its leaves into a cycle. Since a generalized Halin graph needs not to be planar, we investigate the minimum degree condition for a graph to contain it as a spanning subgraph. We show that there exists a positive integer $n_0$ such that any 3-connected graph with $n\ge n_0$ vertices and minimum degree at least $(2n+3)/5$ contains a spanning generalized Halin graph. As an application, the result implies that under the same condition, the graph $G$ contains a wheel-minor of order at least $n/2$. The minimum degree condition in the result is best possible.

math.CO↗

arXiv2014

Characterizing forbidden pairs for hamiltonian squares

The square of a graph is obtained by adding additional edges joining all pair of vertices of distance two in the original graph. Particularly, if $C$ is a hamiltonian cycle of a graph $G$, then the square of $C$ is called a hamiltonian square of $G$. In this paper, we characterize all possible forbidden pairs, which implies the containment of a hamiltonian square, in a 4-connected graph. The connectivity condition is necessary as, except $K_3$ and $K_4$, the square of a cycle is always 4-connected.

math.CO↗

arXiv2014

Vizing's 2-factor Conjecture Involving Large Maximum Degree

Let $G$ be a connected simple graph of order $n$ and let $Δ(G)$ and $χ'(G)$ denote the maximum degree and chromatic index of $G$, respectively. Vizing proved that $χ'(G)=Δ(G)$ or $Δ(G)+1$. Following this result, $G$ is called $Δ$-critical if $χ'(G)=Δ(G)+1$ and $χ'(G-e)=Δ(G)$ for every $e\in E(G)$. In 1968, Vizing conjectured that if $G$ is an $n$-vertex $Δ$-critical graph, then the independence number $α(G)\le n/2$. Furthermore, he conjectured that, in fact, $G$ has a 2-factor. Luo and Zhao showed that if $G$ is an $n$-vertex $Δ$-critical graph with $Δ(G)\ge n/2$, then $α(G)\le n/2$. More recently, they showed that if $G$ is an $n$-vertex $Δ$-critical graph with $Δ(G)\ge 6n/7$, then $G$ has a hamiltonian cycle, and so $G$ has a 2-factor. In this paper, we show that if $G$ is an $n$-vertex $Δ$-critical graph with $Δ(G)\ge n/2$, then $G$ has a 2-factor.

math.CO↗