Assefaw's page on Combinatorial Scientific Computing

A good deal of my research work nicely fits within the field that has come to be known as Combinatorial Scientific Computing (CSC). CSC is an interdisciplinary research area where one pursues and applies tools from combinatorial mathematics and algorithmics to devise efficient methods for computing in science and engineering.

The domain of CSC goes beyond traditional scientific computing (that is, numerical linear algebra, numerical solution of differential equations, and numerical optimization) to include emerging computational sciences, computational engineering, and data science. Prominent among these rapidly evolving disciplines are: computational biology and bioinformatics, computational chemistry and material sciences, computational climate sciences, and statistical physics, among the sciences; simulation-based computational applications in biomedical, chemical, electrical, industrial, and mechanical engineering; and large-scale data management and analysis, across disciplines. Orthogonally, CSC-problems could also emanate from infrastructural technologies supporting high-performance computing. Examples include algorithmic and software tools for parallel processing (such tasks as partitioning and load-balancing), and tools for data and iteration reordering for irregular computations.

Many of the combinatorial problems in the aforementioned seemingly disparate areas are unified by a common feature: they can often be posed and effectively solved as graph problems. Since problems in CSC often are part of a larger computational context, the practical efficiency of their algorithms and the deployment of their implementations via software toolkits is of paramount importance.

The role of combinatorial algorithms in scientific computing has been recognized for several decades, but the banner CSC is coined and the community around it more formally organized fairly recently. The SIAM Workshop on CSC is the premier international event for the field. The first (founding) workshop was organized by John Gilbert, Bruce Hendrickson (Co-Chair), Alex Pothen (Co-Chair), Horst Simon, and Sivan Toledo and was held in 2004 in San Francisco, CA. Since then the workshop has continued to be held regularly once every two years (see list below). As problems and data sets increase in size and complexity, and as parallel computers increase in abundance and sophistication, the importance of CSC will continue to grow in the future.

CSC Workshop Series Select Other CSC Events