Abstract:

Sampling algorithms based on Markov chains arise in many areas of computation, engineering and science. The idea is to perform a random walk among the elements in a large state space so that samples chosen from the stationary distribution are useful for the application. In order to get reliable results efficiently, we require the chain to be rapidly mixing, or quickly converging to equilibrium. Often there is a parameter of the system (typically related to temperature or fugacity) so that at low values many natural chains converge rapidly while at high values they converge slowly, requiring exponential time. This dichotomy is often related to phase transitions in the underlying models. In this talk we will explain this phenomenon, giving examples form the natural and social sciences including magnetization, lattice gasses, colloids, and discrete models of segregation.

This lecture is aimed at a regular audience.