Projects: on the theme of Molecular and Electronic Structure

High Performance Computing for Designing Groundwater Remediation Systems. Mentor: Shawn Matott

Dr. Matott investigates the capabilities and limitations of using predictive environmental models to support policy and management decisions. For example, he has developed uncertainty analysis tools for policy-relevant integrated environmental models used by the US EPA. These tools will aid in the analysis of proposed and ongoing operations involving mountaintop mining, livestock managment, carbon sequestration and hydrofracking. For the URGE 2012 project, Dr. Matott and his cohort will explore the use of parallel optimization algorithms to accelerate the design of groundwater redmediation systems. Much of the research will be performed on the distributed computing cluster hosted by the University at Buffalo's Center for Computational Research (UB CCR). More information here (also see The Project Wiki, Dr. Matott's Home Page and Dr. Matott's Page at the UB CCR).

Mathematical and computational crystallography. Mentor: Hongliang (Jimmy) Xu

Prof. Xu’s research in mathematical and computational crystallography includes (1) the development of novel mathematical models for crystallographic phase problem, (2) the investigation of mathematical optimization procedures for determining three-dimensional crystal structures from X-ray diffraction data. The URGE research cohort will develop and test mathematical optimization algorithms to solve the crystallographic phase problem in a reliable and efficient way. Some of the research activities will take place at the Hauptman-Woodward Institute, a world-class research laboratory in structural and computational biology. More information here.

Evolutionary algorithm for structure prediction of a solid or discrete molecular cluster. Mentor: Eva Zurek.

Structure prediction is an exceedingly difficult task due to the complexity of the potential energy landscape of a solid, or discrete molecular cluster. Simulated annealing, metadynamics and minima hopping are three promising search methods. They are very useful if the system is characterized reasonably well, but if the structure is wholly unknown, genetic or evolutionary algorithms (EAs) may be the best automated search techniques. More information here.