(4bm) Applications of High Performance Computing and Importance Sampling to Problems in Statistical Mechanics

Keys, A. S., University of Michigan

The field of statistical mechanics currently sits at the precipice of major breakthroughs in both its computational and theoretical components.  Over the next decade, cloud computing and general-purpose GPU hardware will revolutionize the way molecular simulations are performed.  Simultaneously, new importance sampling techniques will extend the reach of molecular simulation beyond typical equilibrium thermodynamics to far-from-equilibrium systems often encountered in clean energy, colloidal self-assembly, and biological applications.

In this poster, I present excerpts from my Ph.D. and post-doctoral research in the areas of soft matter and condensed materials.  My Ph.D. work applies high-performance computing and importance sampling techniques to study the self-assembly of nano-and-micro scale target structures.  In particular, my research focuses on assembling materials with a photonic band gap, which allows for the fabrication of tiny switches that can control the passage of light.  My postdoctoral work applies importance sampling of non-equilibrium trajectory space to study the structure and dynamics of glassy materials.  One important question my work addresses is how glasses differ from frozen liquids, and in particular, why a glass reacts very differently than an instantaneously-frozen liquid when melted.  Important unifying concepts throughout my work are the ability of fluctuations and entropy to drive systems into interesting and unexpected structures, and the extension and application of importance sampling techniques to non-equilibrium phenomena.