Dr. Juan C. Meza
Lawrence Berkeley National Laboratory
High Performance Computing Research
1 Cyclotron Road
Berkeley, CA 94720
Phone: +1 510 486 7684
Fax: +1 510 486 4300
DVP term expires December 2012
Dr. Meza has degrees in electrical engineering (B.S. 1978 and Masters 1979, Rice University) and computational mathematics (Masters, Ph.D., 1986, Rice University). He is currently the Department Head of High Performance Computing Research at Lawrence Berkeley National Laboratory. This department focuses on research in scientific data management, visualization, numerical algorithms, and computational sciences and engineering. His current research interests include nonlinear optimization with an emphasis on parallel methods for simulation-based optimization. He has also worked on various scientific and engineering applications including methods for electronic structure calculations for nanoscience applications, combinatorial methods for detecting vulnerabilities in electric power grids, optimization methods for molecular conformation problems, optimal design of chemical vapor deposition furnaces, and semiconductor device modeling. In 2008, Dr. Meza was a member of the winning team in the ACM Gordon Bell Prize competition.
Advanced Algorithms for Analyzing Vulnerabilities in the Electric Power Grid
As the U.S. dependence on networks such as the electric power, communication, and transportation systems grows, the need for secure and reliable operational standards becomes vital to economic, energy and national security. A striking example is the August 2003 Northeast blackout, which demonstrated the catastrophic consequences of a few broken links in a critical infrastructure. In this talk, we will present a screening technology for analyzing vulnerabilities in the electric power grid through a new algorithm that optimally computes the trade-off between the number of suppressed lines and the resulting severity of a blackout. In contrast to other proposed solutions that are either limited to heuristics or rely on supercomputers, our formulation enables us to calculate an optimal solution quickly. In our studies of a 13,374 node power grid model for example, we found optimal solutions within seconds on a desktop computer within minutes. Our tools are applicable to other critical infrastructures where a commodity is distributed from a set of specified sources to terminals.
Audience: 2nd-year graduate students
Surface structure determination of nanostructures using a mesh adaptive optimization method
Many properties of nanostructures depend on the atomic configuration at the surface. One common technique used for determining this surface structure is based on the low energy electron diffraction (LEED) method, which uses a sophisticated physics model to compare experimental results with spectra computed via a computer simulation. While this approach is highly effective, the computational cost of the simulations can be prohibitive for large systems. In this work, we propose the use of generalized pattern search methods, which can handle both discrete and continuous variable and allows the simultaneous optimization of the atomic coordinates as well as the chemical identity.
Audience: 2nd-year graduate students
The Role of Computational Sciences in Amplifying Science Research
Computer modeling and simulation of physical processes has taken on an increasingly larger role in scientific research. In fact, computational science has become what some people term the "third pillar" of science along with theory and experimentation. This increased role is due partly to the tremendous growth in computational power. More importantly, however, the increased role is a direct result of a better understanding of the underlying mathematics and the development of improved algorithms. Examples from wide-ranging fields such as nanoscience, biology, climate modeling and astrophysics point not only to the role that mathematics plays in modeling physical processes but also in predicting new phenomena. In this talk, I will discuss several areas where mathematics has had a profound impact on science, and the role that mathematics has played in “amplifying” the research.