Automated uncertainty quantification for numerical solution of partial differential equations

AUQ-PDE is a collaborative Nordic research project (2015-2019) funded by the Nordforsk/the Nordic Council of Ministers under the Nordic eScience Globalisation Initiative (NeGI). I initiated and led this project from 2015-2017.

The overarching goal of the AUQ-PDE project is to develop and integrate generic software featuring a high degree of automation for uncertainty quantification in computational models governed by partial differential equations and to apply the developed tools to a select set of research questions stemming from the in silico study of physiological processes. To reach this goal, we will:

  • develop novel, high-performance software components, featuring a high degree of automation, for uncertainty quantification, including both forward uncertainty quantification and Bayesian inversion, in computational models for physical processes governed by partial differential equations;
  •  apply these software components to a select set of applications from the biomedical domain, in particular in the modelling of cardiac electrophysiology and mechanics;
  • disseminate these software components as accessible and user-friendly software to the Nordic region and the scientific community in general.

The project brings together three partners from the Nordic countries: the Biomedical Computing Department at Simula Research Laboratory, Norway; the Department of Mathematical Sciences at Chalmers University of Technology, Sweden; and Department of Mathematics and Statistics at the University of Helsinki, Finland.