Photo of LGRTLederle Graduate Research Tower

The Department of Mathematics and Statistics is a community of scholars committed to excellence in research and instruction. We offer a comprehensive set of curricula in our disciplines, from introductory-level general education courses to doctoral dissertation direction and postdoctoral mentoring. Undergraduate majors enjoy a broad array of options through which they can earn the bachelor's degree, and can also apply to participate in summer research activities. The Department's Ph.D. program appears among the top public graduate programs in the recent National Research Council rankings. The M.S. programs in both Applied Mathematics and Statistics contribute to an important pipeline of professionally trained students who enter the high-technology industrial sector.

Faculty News Briefs

September 2016

On September 24, 2016 Visiting Assistant Professor Stathis Charalampidis gave a talk at the Fall Eastern Sectional Meeting of the American Mathematical Society titled "Multi-Component Nonlinear Waves in One and Two Dimensional Coupled Nonlinear Schroedinger Systems: Theory and Numerical Computations". The meeting was held at Bowdoin College in Brunswick, Maine.

A paper, authored by graduate student Konstantinos Gourgoulias, Professor Markos Katsoulakis, and Professor Luc Rey-Bellet, was accepted for publication in the SIAM Journal on Scientific Computing. The paper, titled "Information Metrics for Long-Time Errors in Splitting Schemes for Stochastic Dynamics and Parallel KMC", can be viewed at https://arxiv.org/abs/1511.08240. In this work, the authors study the long-time properties of the parallel Kinetic Monte Carlo algorithm, which is a high-performance computing algorithm used to simulate stochastic models on a lattice with applications as varied as computational physics, traffic modeling, and systems biology. An information criterion connected with path-space relative entropy is derived. Through this information criterion, practitioners can assess the appropriateness of different versions of the algorithm and select the one that is best suited to their application of choice. The proposed methodology extends the numerical analysis of the algorithm to the long-time regime while still providing a tractable computational diagnostic that can be computed during a simulation.

Professor Rob Kusner lectured on "CMC Surfaces and CSC Metrics with All Ends Cylindrical" in the Penn Minimal Surfaces Seminar on September 27, 2016.