Stochastic Loewner evolution driven by Levy processes

Abstract

Standard stochastic Loewner evolution (SLE) is driven by a continuous Brownian motion, which then produces a continuous fractal trace. If jumps are added to the driving function, the trace branches. We consider a generalized SLE driven by a superposition of a Brownian motion and a stable Levy process. The situation is defined by the usual SLE parameter, $κ$, as well as $α$ which defines the shape of the stable Levy distribution. The resulting behavior is characterized by two descriptors: $p$, the probability that the trace self-intersects, and $\tilde{p}$, the probability that it will approach arbitrarily close to doing so. Using Dynkin's formula, these descriptors are shown to change qualitatively and singularly at critical values of $κ$ and $α$. It is reasonable to call such changes ``phase transitions''. These transitions occur as $κ$ passes through four (a well-known result) and as $α$ passes through one (a new result). Numerical simulations are then used to explore the associated touching and near-touching events.