Project Application Details

Dr. Victorita Dolean, Reader, University of Strathclyde Victorita.Dolean@strath.ac.uk

Dr. Niall Bootland, Research Associate, University of Strathclyde Niall.Bootland@strath.ac.uk

Dr. Alexandros Kyriakis, PhD Student, University of Strathclyde Alexandros.Kyriakis@strath.ac.uk

The proposal possess a unique combination of skills that makes this project internationally distinctive, at the interface of Numerical Analysis (NA), Mathematical Analysis (MA) and Scientific Computing (SC). The results we seek to obtain will also impact other disciplines interested in the waves propagation (e.g. electrical engineering, geophysics); the methods described in the Case for Support have the potential to change the numerical simulation of complex problems appearing in these disciplines.
How will the impact be achieved? The first step will be disseminating the results to the research community during the course of the project via
• Publishing papers in the highest quality journals: the investigators will continue to publish their work in top numerical analysis and scientific computing journals.
• Talks in conferences and research visits conducted by the two PDRAs and the investigators. The high-quality publications will lead to presentations at major international meetings (e.g. WAVES, SIAM Geosciences, etc.) and promote the role of UK applied mathematics in this field.
• Software development in open source finite element libraries A suite of examples of discretizations of frequency-domain wave problems and appropriate solvers will be implemented in Freefem++
(http://www.freefem.org/).
• High performance computing (HPC) libraries: the HPC domain decomposition library HPDDM
(https://github.com/hpddm/) will be enhanced so that it contains the solvers developed in this project. This will be freely available as a library to the whole UK community and hence available as a solver for systems generated by any existing finite element code for Helmholtz, Maxwell, or the elastic wave equation. HPDDM will be linked to the Firedrake project (http: //firedrakeproject.org/), so that users can develop code in Python, using the Firedrake
“separation of concerns” philosophy, but also make use of domain decomposition solvers for frequency domain wave problems.
• Annual workshops will bring together interested parties from UK and France. The mains goals of these workshops will be (i) disseminating the results of the project already achieved to the interested scientific community (we plan a mixture of numerical analysts, computational scientists but also electrical engineers and geophysicists), and (ii) exchange with the present experts to benefit the remainder of the project by extending the network of collaborators.
The easy use of these three libraries will make more widely available the methods we develop for a big variety of applications so that UK practitioners can develop applications in Python and have the benefit of the solvers which we develop in this project and we expect that the dissemination of the results will be enhanced by the numerical developments in these libraries.