Computational aspects of uncertainty propagation in large and multiscale systems.

This project aims at the development of numerical algorithms that allow to quantify and control the propagation of parameter uncertainties in mathematical models. We investigate the computational aspects of uncertainty propagation in multiscale models, where the stochasticity of the fine-scale model affects the coarse model and its solution. We examine how to reduce the resulting errors at the coarse scale, and how to match them with spatial and temporal discretization errors. For partial differential equations with uncertain coefficients, we explore new approaches towards the development and analysis of fast and robust iterative solution procedures.

Keywords:Partial differential equations, Uncertainty propagation, Error control, Parameter uncertainties, Fast solvers, Multiscale models, Multigrid method

Disciplines:Applied mathematics in specific fields, Computer architecture and networks, Distributed computing, Information sciences, Information systems, Programming languages, Scientific computing, Theoretical computer science, Visual computing, Other information and computing sciences