SC²S Colloquium - June 10, 2011
|Date:||June 10, 2011|
|Time:||10:00 am, s.t.|
John Jakeman: Uncertainty Quantification Using Sparse Grids
Mathematical models of physical systems are subject to a wide range of uncertainties. The effects of such uncertainty should be traced through the system thoroughly enough to allow one to evaluate their effects on the intended use of the model. Uncertainties occur in all stages of the modelling process and assume many forms. In this talk I will present sparse grid based methods for quantifying input uncertainty driven by stochastic processes and/or arising from unknown model parameters. Specifically I will discuss the role of sparse grids for statistical moment estimation, sensitivity analysis, reliability based design and Bayesian inference.