Seminar WS 21/22 Graduate Seminar on Scientific Computing
Uncertainty Quantification
Approximate solutions of partial differential equations computed by numerical methods can nowadays often be calculated with sufficient accuracy. The basic assumption of these methods is that the underlying input data are known exactly. In practice, however, this is often not the case, and the input data of the partial differential equations are subject to uncertainties. As a consequence, solutions calculated using this input data are also subject to uncertainties, so that an exact approximate solution of these equations is at least questionable.In this seminar, we will look at various techniques for quantifying the uncertainties in solutions of partial differential equations with uncertain input data.