Searcharxiv⌕ Search

arXiv subjects

Allan Lo

Publications and source records attributed to Allan Lo.

66 records · Page 4Linked to original sources

Perfect matchings in 3-partite 3-uniform hypergraphs

Let $H$ be a $3$-partite $3$-uniform hypergraph, i.e. a $3$-uniform hypergraph such that every edge intersects every partition class in exactly one vertex, with each partition class of size $n$. We determine a Dirac-type vertex degree threshold for perfect matchings in $3$-partite $3$-uniform hypergraphs.

math.CO↗

$\ell$-degree Turán density

Let $H_n$ be a $k$-graph on $n$ vertices. For $0 \le \ell \ell >1$, the set of $π_{\ell}^k(\mathcal{F})$ is dense in the interval $[0,1)$. Hence, there is no "jump" for $\ell$-degree Turán density when $k>\ell >1$. We also give a lower bound on $π_{\ell}^k(\mathcal{F})$ in terms of an ordinary Turán density.

math.CO↗

Properly coloured Hamiltonian cycles in edge-coloured complete graphs

Let $K_n^c$ be an edge-coloured complete graph on $n$ vertices. Let $Δ_{\rm mon}(K_n^c)$ denote the largest number of edges of the same colour incident with a vertex of $K_n^c$. A properly coloured cycle is a cycle such that no two adjacent edges have the same colour. In 1976, Bollobás and Erdős conjectured that every $K_n^c$ with $Δ_{\rm mon}(K_n^c) < \lfloor n/2 \rfloor$ contains a properly coloured Hamiltonian cycle. In this paper, we show that for any $\varepsilon > 0 $, there exists an integer $n_0$ such that every $K_n^c$ with $Δ_{\rm mon}(K_n^c) < (1/2 - \varepsilon) n $ and $n \ge n_0$ contains a properly coloured Hamiltonian cycle. This improves a result of Alon and Gutin. Hence, the conjecture of Bollobás and Erdős is true asymptotically.

math.CO↗

The robust component structure of dense regular graphs and applications

In this paper, we study the large-scale structure of dense regular graphs. This involves the notion of robust expansion, a recent concept which has already been used successfully to settle several longstanding problems. Roughly speaking, a graph is robustly expanding if it still expands after the deletion of a small fraction of its vertices and edges. Our main result allows us to harness the useful consequences of robust expansion even if the graph itself is not a robust expander. It states that every dense regular graph can be partitioned into `robust components', each of which is a robust expander or a bipartite robust expander. We apply our result to obtain (amongst others) the following. (i) We prove that whenever $\eps >0$, every sufficiently large 3-connected D-regular graph on n vertices with $D \geq (1/4 + \eps)n$ is Hamiltonian. This asymptotically confirms the only remaining case of a conjecture raised independently by Bollobás and Häggkvist in the 1970s. (ii) We prove an asymptotically best possible result on the circumference of dense regular graphs of given connectivity. The 2-connected case of this was conjectured by Bondy and proved by Wei.

math.CO↗

$F$-factors in hypergraphs via absorption

Given integers $ n \ge k >l \ge 1 $ and a $k$-graph $F$ with $|V(F)|$ divisible by $n$, define $t_l^k(n,F)$ to be the smallest integer $d$ such that every $k$-graph $H$ of order $n$ with minimum $l$-degree $δ_l(H) \ge d $ contains an $F$-factor. A classical theorem of Hajnal and Szemerédi implies that $t^2_1(n,K_t) = (1-1/t)n$ for integers $t$. For $k \ge 3$, $t^k_{k-1}(n,K_k^k)$ (the $δ_{k-1}(H)$ threshold for perfect matchings) has been determined by Kühn and Osthus (asymptotically) and Rödl, Ruciński and Szemerédi (exactly) for large $n$. In this paper, we generalise the absorption technique of Rödl, Ruciński and Szemerédi to $F$-factors. We determine the asymptotic values of $t^k_1(n,K_k^k(m))$ for $k = 3,4$ and $m \ge 1$. In addition, we show that for $t>k = 3$ and $γ>0$, $ t^3_{2}(n,K_t^3) \le (1- \frac{2}{t^2-3t+4} + γ) n$ provided $n$ is large and $t | n$. We also bound $t^3_{2}(n,K_t^3)$ from below. In particular, we deduce that $t^3_2(n,K_4^3) = (3/4+o(1))n$ answering a question of Pikhurko. In addition, we prove that $t^k_{k-1}(n,K_t^k) \le (1- \binom{t-1}{k-1}^{-1} + γ)n$ for $γ>0$, $k \ge 6$ and $t \ge (3+ \sqrt5)k/2$ provided $n$ is large and $t | n$.

math.CO↗

An edge-coloured version of Dirac's theorem

Let $G$ be an edge-coloured graph. The minimum colour degree $ δ^c(G) $ of $G$ is the largest integer $k$ such that, for every vertex $v$, there are at least $k$ distinct colours on edges incident to $v$. We say that $G$ is properly coloured if no two adjacent edges have the same colour. In this paper, we show that every edge-coloured graph $G$ with $ δ^c(G) \ge 2|G| / 3 $ contains a properly coloured $2$-factor. Furthermore, we show that for any $ \varepsilon > 0 $ there exists an integer $ n_0 $ such that every edge-coloured graph $G$ with $|G| = n \ge n_0 $ and $ δ^c(G) \ge ( 2/3 + \varepsilon ) n $ contains a properly coloured cycle of length $\ell$ for every $3 \le \ell \le n$. This result is best possible in the sense that the statement is false for $ δ^c(G) < 2n / 3 $.

math.CO↗

A Dirac type condition for properly coloured paths and cycles

Let $c$ be an edge-colouring of a graph $G$ such that for every vertex $v$ there are at least $d \ge 2$ different colours on edges incident to $v$. We prove that $G$ contains a properly coloured path of length 2d or a properly coloured cycle of length at least $d+1$. Moreover, if $G$ does not contain any properly coloured cycle, then there exists a properly coloured path of length $3 \times 2^{d-1}-2$.

math.CO↗

A multipartite version of the Hajnal-Szemerédi theorem for graphs and hypergraphs

A perfect $K_t$-matching in a graph $G$ is a spanning subgraph consisting of vertex disjoint copies of $K_t$. A classic theorem of Hajnal and Szemerédi states that if $G$ is a graph of order $n$ with minimum degree $δ(G) \ge (t-1)n/t$ and $t| n$, then $G$ contains a perfect $K_t$-matching. Let $G$ be a $t$-partite graph with vertex classes $V_1$,..., $V_t$ each of size $n$. We show that if every vertex $x \in V_i$ is joined to at least $((t-1)/t + γ)n $ vertices of $V_j$ for $i \ne j$, then $G$ contains a perfect $K_t$-matching, thus verifying a conjecture of Fisher asymptotically. Furthermore, we consider a generalisation to hypergraphs in terms of the codegree.

math.CO↗

Minimum codegree threshold for $(K_4^3-e)$-factors

Given hypergraphs H and F, an F-factor in H is a spanning subgraph consisting of vertex disjoint copies of F. Let K_4^3-e denote the 3-uniform hypergraph on 4 vertices with 3 edges. We show that for γ>0 there exists an integer n_0 such that every 3-uniform hypergraph $H$ of order n > n_0 with minimum codegree at least (1/2+γ)n and 4|n contains a (K_4^3-e)-factor. Moreover, this bound is asymptotically the best possible and we further give a conjecture on the exact value of the threshold for the existence of a (K_4^3-e)-factor. Therefore, all minimum codegree thresholds for the existence of F-factors are known asymptotically for 3-uniform hypergraphs F on 4 vertices.

math.CO↗

A note on large rainbow matchings in edge-coloured graphs

A rainbow subgraph in an edge-coloured graph is a subgraph such that its edges have distinct colours. The minimum colour degree of a graph is the smallest number of distinct colours on the edges incident with a vertex over all vertices. Kostochka, Pfender, and Yancey showed that every edge-coloured graph on $n$ vertices with minimum colour degree at least $k$ contains a rainbow matching of size at least $k$, provided $n\geq (17/4)k^2$. In this paper, we show that $n\geq 4k-4$ is sufficient for $k \ge 4$.

math.CO↗

A note on rainbow matchings in properly edge-coloured graphs

A rainbow matching in an edge-coloured graph is a matching such that its edges have distinct colours. We show that every properly edge-coloured graph $G$ with $|G| \ge (9δ(G) -5)/2$ has a rainbow matching of size $δ(G)$, improving a result of Diemunsch et al.

math.CO↗

Cliques in graphs with bounded minimum degree

Let $k_r(n,δ)$ be the minimum number of $r$-cliques in graphs with $n$ vertices and minimum degree $δ$. We evaluate $k_r(n,δ)$ for $δ\leq 4n/5$ and some other cases. Moreover, we give a construction, which we conjecture to give all extremal graphs (subject to certain conditions on $n$, $δ$ and $r$).

math.CO↗