Scientific Computing


Scientific Computing, an inherently multidisciplinary research area, emphasizes the development of highly parallel scientific computing applications targeting diverse exascale architectures. The focus is on increasing the understanding of fundamental problems in physical, chemical, and biological sciences and engineering.

The CICS program in scientific computation and numerical analysis is integrated with developments in uncertainty quantification, stochastic optimization, and machine learning, as well as with other research activities in the Center. Special emphasis is given to high-order techniques for the solution of nonlinear PDEs that arise in control theory, fluid dynamics and computational physics, multiscale, multigrid, and multiphysics approximations. Numerical methods for the discontinuous Galerkin methods that arise in shock wave propagation, collocation, and stochastic approximations theories for stochastic PDEs and uncertainty modeling are being studied. Emphasis is also being placed on the use of parallel processors and graphics processing units in large-scale linear and nonlinear problems, compressive sensing, artificial intelligence, machine learning, and deep learning. 

Faculty devoted to this research include Dinshaw BalsaraBei Hu, Karel Matous, Joannes WesterinkNicholas Zabaras, and others.