Event Type:

Applied Mathematics and Computation Seminar

Date/Time:

Friday, November 18, 2016 - 12:00 to 13:00

Location:

GLK 115

Event Link:

Local Speaker:

Abstract:

There are several reasons why a stochastic parametrization is a reasonable option in modeling phenomena: in order to include effects that are poorly understood yet recognized as important to the fidelity of a model when compared to field data; when dimension reduction is essential to the computational implementation of a model; when a model needs to improve with regard to sensitive consistency to parameters.

I will demonstrate how stochastic parametrization was able to improve the fidelity of a longshore current model, how it is being used to model wave breaking dissipation, and in glassy dynamics, which are high-dimensional systems where equilibrium statistical mechanics ideas fail. I will also introduce the notion of consistent sensitivity and how stochasticity can improve models that are rich in parameters, as is often the case in models for biological dynamics.