Financial Engineering Seminar
April 3, 2009 (Fri)
Time: 2:00 PM
Place: Weil Hall 303
The propagation of epistemic and aleatory uncertainty through systems models is difficult, especially when an interval description of uncertainty is given and system models are computationally expensive to evaluate. There have been many probability-based methods for aleatory uncertainty propagation which employ analytical approximations for probability integration, yet how these methods should be applied in the presence of both aleatory and epistemic uncertainty is not clear. This presentation gives a brief overview of analytical methods for uncertainty propagation and probability bounds analysis. It shows how bounds on probability distributions can be constructed for sparse and interval data using flexible families of four-parameter probability distributions, and uses probabilistic methods to propagate uncertainty through systems models. Extensions to epidemic modeling, aerospace drag prediction, and multidisciplinary analysis are discussed.