Statistical features of large fluctuations in stochastic systems

The fluctuation of transitional paths that describe the decay from a metastable state are studied with a space-time Monte Carlo algorithm. For a bistable stochastic system we characterize the statistical properties that describe the growth of large fluctuations. Interesting statistical features are discussed. In particular, we study the fluctuation enhancement (large non-Gaussian fluctuation) that occurs as the system escapes from a metastable state. The results are discussed in the context of a scaling theory for the decay from an unstable state.