Applied and Computational Math Seminar: Computational methods for the study of rare events

Speaker:  Maria Cameron, University of Maryland

Title:  Computational methods for the study of rare events

Abstract:  Many processes in the real world are modeled by stochastic differential equations (SDEs) with small white noise. No matter how small the noise is, such systems can perform transitions between basins of attractors of the corresponding ordinary differential equations on large enough timescales. The quantification of these transitions via direct simulations is difficult due to long waiting times. I will give an overview of deterministic approaches for quantifying rare transitions and then focus on two of them. The first approach based on diffusion maps with Mahalanobis kernel is developed in the context of chemical physics applications. The second approach based on the large deviations theory and the optimal control theory is designed to quantify transitions between various oscillatory modes in noise-controlled mechanical systems with periodic forcing.

Time:  Monday, March 11, 2022, 1:30pm-2:30pm

Place:  In-person  (Exploratory Hall, Room 4106)  and Zoom