The study of random combinatorial objects, such as random graphs, is an attempt to discover what a "typical" graph with certain properties might look like (for instance, a given number of vertices and edges, or a given number of vertices and every vertex having a fixed degree). This work involves a mixture of combinatorial and probabilistic arguments, and the results obtained are asymptotic, meaning that they hold in the limit as the number of vertices of the graph tends to infinity.

A related area is asymptotic enumeration of various kinds of graphs. This means finding a formula for the number of graphs of interest which is closer and closer to the truth as the number of vertices tends to infinity. For sparse graphs the methods are combinatorial and often involve a probabilistic approach using switchings. For dense graphs the approach is quite different: we want to to know a particular coefficient of the generating function and proceed using complex analysis, evaluating integrals using the saddle-point method.

I am also interested in the design and analysis of randomized algorithms for graphs and other combinatorial structures, particularly Markov chain Monte Carlo algorithms for approximate sampling and/or counting. Here a Markov chain is designed with a given stationary distribution, for instance the uniform distribution on the set of all graphs with a given degree sequence. The aim is to prove that the Markov chain converges rapidly (i.e. in polynomial time) to this distribution.