Tuesday, February 28, 2012
101 Stanley Thomas Hall
Tulane University (Uptown)
Refreshments will be served
Kyle Hickmann, Center for Computational Science, Tulane University
Uncertainty Quantification for Epidemic Dynamics of Stochastic Simulations
It is standard practice to run complex simulations of many natural phenomenon using mathematical models based on small scale properties of the system. However, when this is done the output of the simulation may be as difficult to understand as the original observed phenomenon. Researchers have developed uncertainty quantification methods to statistically describe the effect of parameter inputs on simulation output to analyze these models. In the beginning of this talk I will describe some of the common methods used for uncertainty quantification and sensitivity analysis. The second part of this talk will focus on uncertainty quantification for EpiSimS, an agent based model of epidemic spread through a large population in use at Los Alamos National Laboratory. I will describe tools developed to study the dynamics of disease spread using stochastic simulations. The methods focus on finding a set of variables that characterize the deterministic and stochastic parts of the disease progression separately. These variables may then be used for a sensitivity analysis and uncertainty quantification study.
Center for Computational Science, Stanley Thomas Hall 402, New Orleans, LA 70118 email@example.com