Worm Monte Carlo dynamics for Loop models and Potts antiferromagnets
by Dr Tim Garoni
Abstract: In this talk I'll describe a "worm" dynamics (Markov-chain Monte Carlo
algorithm) for the O(n)-loop model on bipartite cubic graphs for general n>0. In particular, I'll show that this new dynamics remains ergodic in the fully-packed limit. As an application, I'll show how this allows us to simulate certain zero-temperature antiferromagnetic Potts models (aka graph-colouring problems), for which cluster algorithms have recently been proved to fail. This talk will be largely theoretical in nature, and no numerical results will be discussed.
For More Information: contact: Mark Sorrell. email: firstname.lastname@example.org