arXiv subjects

Colin McDiarmid

Publications and source records attributed to Colin McDiarmid.

At least 37 records · Page 2Linked to original sources

arXiv2017

Random Perfect Graphs

We investigate the asymptotic structure of a random perfect graph $P_n$ sampled uniformly from the perfect graphs on vertex set $\{1,\ldots,n\}$. Our approach is based on the result of Prömel and Steger that almost all perfect graphs are generalised split graphs, together with a method to generate such graphs almost uniformly. We show that the distribution of the maximum of the stability number $α(P_n)$ and clique number $ω(P_n)$ is close to a concentrated distribution $L(n)$ which plays an important role in our generation method. We also prove that the probability that $P_n$ contains any given graph $H$ as an induced subgraph is asymptotically $0$ or $\frac12$ or $1$. Further we show that almost all perfect graphs are $2$-clique-colourable, improving a result of Bacsó et al from 2004; they are almost all Hamiltonian; they almost all have connectivity $κ(P_n)$ equal to their minimum degree; they are almost all in class one (edge-colourable using $Δ$ colours, where $Δ$ is the maximum degree); and a sequence of independently and uniformly sampled perfect graphs of increasing size converges almost surely to the graphon $W_P(x, y) = \frac12(\mathbb{1}[x \le 1/2] + \mathbb{1}[y \le 1/2])$.

math.CO↗

arXiv2017

Modularity of regular and treelike graphs

Clustering algorithms for large networks typically use modularity values to test which partitions of the vertex set better represent structure in the data. The modularity of a graph is the maximum modularity of a partition. We consider the modularity of two kinds of graphs. For $r$-regular graphs with a given number of vertices, we investigate the minimum possible modularity, the typical modularity, and the maximum possible modularity. In particular, we see that for random cubic graphs the modularity is usually in the interval $(0.666, 0.804)$, and for random $r$-regular graphs with large $r$ it usually is of order $1/\sqrt{r}$. These results help to establish baselines for statistical tests on regular graphs. The modularity of cycles and low degree trees is known to be close to 1: we extend these results to `treelike' graphs, where the product of treewidth and maximum degree is much less than the number of edges. This yields for example the (deterministic) lower bound $0.666$ mentioned above on the modularity of random cubic graphs.

math.CO↗

arXiv2017

List Colouring Squares of Planar Graphs

In 1977, Wegner conjectured that the chromatic number of the square of every planar graph $G$ with maximum degree $Δ\ge8$ is at most $\bigl\lfloor\frac32Δ\bigr\rfloor+1$. We show that it is at most $\frac32 Δ(1+o(1))$ (where the $o(1)$ is as $Δ\to+\infty$), and indeed that this is true for the list chromatic number and for more general classes of graphs.

math.CO↗

arXiv2016

Clique colouring of binomial random graphs

A clique colouring of a graph is a colouring of the vertices so that no maximal clique is monochromatic (ignoring isolated vertices). The smallest number of colours in such a colouring is the clique chromatic number. In this paper, we study the asymptotic behaviour of the clique chromatic number of the random graph G(n,p) for a wide range of edge-probabilities p=p(n). We see that the typical clique chromatic number, as a function of the average degree, forms an intriguing step function.

math.PR↗

arXiv2016

Edge-Maximal Graphs on Surfaces

We prove that for every surface $Σ$ of Euler genus $g$, every edge-maximal embedding of a graph in $Σ$ is at most $O(g)$ edges short of a triangulation of $Σ$. This provides the first answer to an open problem of Kainen (1974).

math.CO↗

arXiv2016

Random graphs from a block-stable class

A class of graphs is called block-stable when a graph is in the class if and only if each of its blocks is. We show that, as for trees, for most $n$-vertex graphs in such a class, each vertex is in at most $(1+o(1)) \log n / \log\log n$ blocks, and each path passes through at most $5 (n \log n)^{1/2}$ blocks. These results extend to `weakly block-stable' classes of graphs.

math.CO↗

arXiv2016

Hamilton cycles, minimum degree and bipartite holes

We present a tight extremal threshold for the existence of Hamilton cycles in graphs with large minimum degree and without a large ``bipartite hole`` (two disjoint sets of vertices with no edges between them). This result extends Dirac's classical theorem, and is related to a theorem of Chvátal and Erdős. In detail, an $(s, t)$-bipartite-hole in a graph $G$ consists of two disjoint sets of vertices $S$ and $T$ with $|S|= s$ and $|T|=t$ such that there are no edges between $S$ and $T$; and $\widetildeα(G)$ is the maximum integer $r$ such that $G$ contains an $(s, t)$-bipartite-hole for every pair of non-negative integers $s$ and $t$ with $s + t = r$. Our central theorem is that a graph $G$ with at least $3$ vertices is Hamiltonian if its minimum degree is at least $\widetildeα(G)$. From the proof we obtain a polynomial time algorithm that either finds a Hamilton cycle or a large bipartite hole. The theorem also yields a condition for the existence of $k$ edge-disjoint Hamilton cycles. We see that for dense random graphs $G(n,p)$, the probability of failing to contain many edge-disjoint Hamilton cycles is $(1 - p)^{(1 + o(1))n}$. Finally, we discuss the complexity of calculating and approximating $\widetildeα(G)$.

math.CO↗

arXiv2016

Recognition of unipolar and generalised split graphs

A graph is unipolar if it can be partitioned into a clique and a disjoint union of cliques, and a graph is a generalised split graph if it or its complement is unipolar. A unipolar partition of a graph can be used to find efficiently the clique number, the stability number, the chromatic number, and to solve other problems that are hard for general graphs. We present the first $O(n^2)$ time algorithm for recognition of $n$-vertex unipolar and generalised split graphs, improving on previous $O(n^3)$ time algorithms.

cs.DS↗

arXiv2016

Uniform multicommodity flow in the hypercube with random edge capacities

We give two results for multicommodity flows in the $d$-dimensional hypercube ${Q}^d$ with independent random edge capacities distributed like $C$ where $\Pr[C>0]>1/2$. Firstly, with high probability as $d \rightarrow \infty$, the network can support simultaneous multicommodity flows of volume close to $E[C]$ between all antipodal vertex pairs. Secondly, with high probability, the network can support simultaneous multicommodity flows of volume close to $2^{1-d} E[C]$ between all vertex pairs. Both results are best possible.

math.OC↗

arXiv2016

Connectivity for bridge-alterable graph classes

A collection of graphs is called bridge-alterable if, for each graph G with a bridge e, G is in the class if and only if G-e is. For example the class of forests is bridge-alterable. For a random forest $F_n$ sampled uniformly from the set of forests on vertex set {1,..,n}, a classical result of Renyi (1959) shows that the probability that $F_n$ is connected is $e^{-1/2 +o(1)}$. Recently Addario-Berry, McDiarmid and Reed (2012) and Kang and Panagiotou (2013) independently proved that, given a bridge-alterable class, for a random graph $R_n$ sampled uniformly from the graphs in the class on {1,..,n}, the probability that $R_n$ is connected is at least $e^{-1/2 +o(1)}$. Here we give a more straightforward proof, and obtain a stronger non-asymptotic form of this result, which compares the probability to that for a random forest. We see that the probability that $R_n$ is connected is at least the minimum over $\frac25 n < t \leq n$ of the probability that $F_t$ is connected.

math.CO↗

arXiv2015

Extremal Distances for Subtree Transfer Operations in Binary Trees

Three standard subtree transfer operations for binary trees, used in particular for phylogenetic trees, are: tree bisection and reconnection ($TBR$), subtree prune and regraft ($SPR$) and rooted subtree prune and regraft ($rSPR$). For a pair of leaf-labelled binary trees with $n$ leaves, the maximum number of such moves required to transform one into the other is $n-Θ(\sqrt{n})$, extending a result of Ding, Grunewald and Humphries. We show that if the pair is chosen uniformly at random, then the expected number of moves required to transfer one into the other is $n-Θ(n^{2/3})$. These results may be phrased in terms of agreement forests: we also give extensions for more than two binary trees.

math.CO↗

arXiv2015

Balanced routing of random calls

We consider an online network routing problem in continuous time, where calls have Poisson arrivals and exponential durations. The first-fit dynamic alternative routing algorithm sequentially selects up to $d$ random two-link routes between the two endpoints of a call, via an intermediate node, and assigns the call to the first route with spare capacity on each link, if there is such a route. The balanced dynamic alternative routing algorithm simultaneously selects $d$ random two-link routes, and the call is accepted on a route minimising the maximum of the loads on its two links, provided neither of these two links is saturated. We determine the capacities needed for these algorithms to route calls successfully and find that the balanced algorithm requires a much smaller capacity. In order to handle such interacting random processes on networks, we develop appropriate tools such as lemmas on biased random walks.

math.PR↗

arXiv2012

Random graphs from a weighted minor-closed class

There has been much recent interest in random graphs sampled uniformly from the n-vertex graphs in a suitable minor-closed class, such as the class of all planar graphs. Here we use combinatorial and probabilistic methods to investigate a more general model. We consider random graphs from a `well-behaved' class of graphs: examples of such classes include all minor-closed classes of graphs with 2-connected excluded minors (such as forests, series-parallel graphs and planar graphs), the class of graphs embeddable on any given surface, and the class of graphs with at most k vertex-disjoint cycles. Also, we give weights to edges and components to specify probabilities, so that our random graphs correspond to the random cluster model, appropriately conditioned. We find that earlier results extend naturally in both directions, to general well-behaved classes of graphs, and to the weighted framework, for example results concerning the probability of a random graph being connected; and we also give results on the 2-core which are new even for the uniform (unweighted) case.

math.CO↗

arXiv2012

Independent sets in graphs with given minimum degree

We consider numbers and sizes of independent sets in graphs with minimum degree at least $d$, when the number $n$ of vertices is large. In particular we investigate which of these graphs yield the maximum numbers of independent sets of different sizes, and which yield the largest random independent sets. We establish a strengthened form of a conjecture of Galvin concerning the first of these topics.

math.CO↗

arXiv2012

On the spread of random graphs

The spread of a connected graph G was introduced by Alon, Boppana and Spencer (1998) and measures how tightly connected the graph is. It is defined as the maximum over all Lipschitz functions f on V(G) of the variance of f(X) when X is uniformly distributed on V(G). We investigate the spread for certain models of sparse random graph; in particular for random regular graphs G(n,d), for Erdős-Rényi random graphs G_{n,p} in the supercritical range p>1/n, and for a 'small world' model. For supercritical G_{n,p}, we show that if p=c/n with c>1 fixed then with high probability the spread of the giant component is bounded, and we prove corresponding statements for other models of random graphs, including a model with random edge-lengths. We also give lower bounds on the spread for the barely supercritical case when p=(1+o(1))/n. Further, we show that for d large, with high probability the spread of G(n,d) becomes arbitrarily close to that of the complete graph K_n.

math.PR↗

arXiv2012

Connectivity for random graphs from a weighted bridge-addable class

There has been much recent interest in random graphs sampled uniformly from the n-vertex graphs in a suitable structured class, such as the class of all planar graphs. Here we consider a general 'bridge-addable' class of graphs - if a graph is in the class and u and v are vertices in different components then the graph obtained by adding an edge (bridge) between u and v must also be in the class. Various bounds are known concerning the probability of a random graph from such a class being connected or having many components, sometimes under the additional assumption that bridges can be deleted as well as added. Here we improve or amplify or generalise these bounds. For example, we see that the expected number of vertices left when we remove a largest component is less than 2. The generalisation is to consider 'weighted' random graphs, sampled from a suitable more general distribution, where the focus is on the bridges.

math.CO↗

arXiv2012

Largest sparse subgraphs of random graphs

For the Erdős-Rényi random graph G(n,p), we give a precise asymptotic formula for the size of a largest vertex subset in G(n,p) that induces a subgraph with average degree at most t, provided that p = p(n) is not too small and t = t(n) is not too large. In the case of fixed t and p, we find that this value is asymptotically almost surely concentrated on at most two explicitly given points. This generalises a result on the independence number of random graphs. For both the upper and lower bounds, we rely on large deviations inequalities for the binomial distribution.

math.CO↗

arXiv2011

Integer realizations of disk and segment graphs

A disk graph is the intersection graph of disks in the plane, a unit disk graph is the intersection graph of same radius disks in the plane, and a segment graph is an intersection graph of line segments in the plane. It can be seen that every disk graph can be realized by disks with centers on the integer grid and with integer radii; and similarly every unit disk graph can be realized by disks with centers on the integer grid and equal (integer) radius; and every segment graph can be realized by segments whose endpoints lie on the integer grid. Here we show that there exist disk graphs on $n$ vertices such that in every realization by integer disks at least one coordinate or radius is $2^{2^{Ω(n)}}$ and on the other hand every disk graph can be realized by disks with integer coordinates and radii that are at most $2^{2^{O(n)}}$; and we show the analogous results for unit disk graphs and segment graphs. For (unit) disk graphs this answers a question of Spinrad, and for segment graphs this improves over a previous result by Kratochv\'ıl and Matou{š}ek.

math.MG↗