A unified framework for umbrella sampling in reversible and irreversible systems

Aaron Dinner
University of Chicago

Umbrella sampling is a means of stratified sampling in molecular simulations. I will describe a mathematical framework that casts umbrella sampling as an eigenvalue problem. This provides a unified perspective on umbrella sampling in both reversible and irreversible systems. To demonstrate the demonstrate the power of the framework, I will show examples in which the strata are defined in terms of both points in time and path-dependent variables.

Presentation (PDF File)

Back to Workshop III: Surrogate Models and Coarsening Techniques