The predictive capabilities of mathematical models (widely used today in engineering, environmental, medical and economical sciences) is strictly connected to how accurate our knowledge of the parameters of such models is. These parameters are however often affected by a large degree of uncertainty, due to a number of reasons, such as measurement errors (or the impracticability of an exhaustive experimental campaign), limited knowledge of the phenomenon under study, or intrinsic randomness of the parameters. The parameters can therefore be described as random variables/fields, and the goal of uncertainty quantification analyses (UQ) is to assess the impact of such uncertainties on the solutions of the mathematical models. Other related problems are optimization under uncertainty, inverse problems and optimal design of experiments.

Uncertainty Quantification is therefore a highly interdisciplinary field, that combines aspects of numerical analysis, scientific computing, mathematical modeling and probability/statistics. Its practical applicability is also significantly connected to the availability of high-performance computing hardware. The aim of this workshop is to discuss the modeling and numerical/methodological aspects of Uncertainty Quantification, and to compare the experiences in this field among DIITET researchers, which are active in a broad spectrum of scientific areas. Some internationally recognized experts will also participate to the workshop, to deliver keynotes on selected topics of interest.