Per-Gunnar Martinsson, University of Oxford.
The 2017 Germund Dahlquist Prize was awarded to Per-Gunnar Martinsson at the
2017 International Conference on Scientific Computation and Differential Equations, held September 11-15, 2017 at the University of Bath, U.K. The Germund Dahlquist Prize is awarded for original contributions to fields associated with Swedish mathematician Germund Dahlquist, especially the numerical solution of differential equations and numerical methods for scientific computing.
“The pioneering work of Germund Dahlquist has been an inspiration to me for a long time,” Martinsson said. “He played a key role in establishing numerical analysis as an area of particular research excellence in Sweden, and I was very fortunate to have the opportunity to start my studies in applied mathematics in that environment.”
The prize honors Martinsson for fundamental contributions to numerical analysis and scientific computing that are having a significant impact in data science applications. The prize committee cites two of Martinsson’s contributions in particular: (1) the development of linear time algorithms for dense matrix operations related to multidimensional elliptic partial differential equations and integral equations, and (2) deep and innovative contributions to the development of probabilistic algorithms for the rapid solution of certain classes of large-scale linear algebra problems.
“The speed of any computational task depends in part on how fast the hardware is and in part on how effective our algorithms are,” Martinsson said. “Getting faster algorithms for solving common linear algebraic tasks means that we can now perform much more accurate computational simulations of physical phenomena since we can include more details and work with more realistic models. The computational tasks on which we have been focusing are also core building blocks in machine learning and data mining, so this has led to better search algorithms, smarter smartphones, faster algorithms for genomics, etc.”
Martinsson is currently Professor of Numerical Analysis at the University of Oxford. He received his Ph.D. in computational and applied mathematics in 2002 from the Institute for Computational Engineering and Sciences at the University of Texas at Austin.
“I am very enthusiastic about the new methods that my collaborators and I have developed over the last several years,” Martinsson said. “Besides providing new tools for computational scientists, I hope that these results have opened up possibilities for new exciting research at the interfaces between mathematics, numerics, and applications. If this award helps to draw attention to randomized methods and fast direct solvers, then that would be very exciting.”