Automated uncertainty quantification for numerical solutions of partial differential equations
Research Project, 2015 – 2017

Uncertainties and measurement errors permeate all fields of computational science. In the biomedical disciplines, the uncertainties inherent in data acquisition and processing pose a fundamental challenge in our era of patient-specific modelling and simulation. The quantification of such uncertainties and their implications is vital for the predictive capabilities of computer simulations. In spite of its importance, the role of uncertainty quantification is yet underdeveloped in these data-driven scientific fields. This can be attributed to a critical and problematic gap between clinicians, biomedical engineers, numerical method and algorithm designers, and scientific software developers. Such gaps are not restricted to the biomedical domain: indeed, these gaps present a generic challenge in the field of computational science.

The AUQ-PDE project aims to bridge this gap by developing and integrating generic software components featuring a high degree of automation for uncertainty quantification in computational models governed by partial differential equations (PDEs). The applicability and usability of the software will be anchored in the application domains by the central involvement of application domain specialists. In particular, the software components will be demonstrated on a select set of research questions stemming from the in silico study of cardiac electrophysiology and mechanics. The software developed will allow scientists and engineers to quickly build tailored PDE models and quickly equip these models with tailored uncertainty quantification methods. In the longer-term, the more widespread availability of biomedical simulation studies with quantified uncertainties will positively impact the use of such studies for clinical practice. For instance, the cardiac modelling study proposed in this project will strengthen the possibility of using patient-specific simulations as a diagnostic or treatment planning tool for cardiac diseases.


Anders Logg (contact)

Chalmers, Mathematical Sciences, Applied Mathematics and Statistics


Simula Research Laboratory

Snaroya, Norway

University of Helsinki

Helsinki, Finland



Project ID: 74756
Funding Chalmers participation during 2015–2017

More information

Latest update