Scientific computing harnesses computing to solve problems in science and engineering that involve mathematical models, such as those based on partial differential equations or the interactions between particles such as individual atoms. The applications range in scale from biomolecules, complex materials, built structures and other engineered systems, weather and climate, to galaxies. Simulation of natural and engineered systems with these models on a computer is particularly important when these systems are inaccessible to actual experiments, e.g., observing proteins as they bind to drug molecules, predicting the landfall of hurricanes or the spread of disease, and mapping the evolution of the universe.
In CSE, the goal of our research in scientific computing is to develop methodologies that enable new science and new engineering that could not have been done before. Our research efforts span all areas of scientific computing, from developing mathematical models, developing numerical and combinatorial algorithms to evaluate or solve these models, and designing and implementing algorithms that run efficiently on highly parallel computers.
Increasingly, data is being used in scientific computing to augment the mathematical models, which is essential when models are inexact or too complex in their exact forms. Scientific computing also involves developing new machine learning approaches to address scientific and engineering problems, for example, multiscale models, data assimilation, uncertainty quantification, high-dimensional sampling, and numerical optimization.
At Georgia Tech, virtually every discipline uses scientific computing in its research, as computation and data-driven analysis have become indispensable and complementary approaches to experimentation and physical prototyping for scientific and engineering discovery. CSE leads or collaborates with other Schools at Georgia Tech and outside of Georgia Tech on projects in chemistry, biology, health, geophysics, and materials science. Scientific computing research at CSE is also closely tied to research efforts in high-performance computing and machine learning.
Center for High Performance Computing (CHiPC)
Institute for Data Engineering and Science (IDEaS)