Michael Mascagni

(732) 601-4470

Research

My research group focuses on all aspects of stochastic computing. We are interested in the development of Monte Carlo and Quasi-Monte Carlo algorithms; the applications of these algorithms in materials science, biochemistry, and physics; and the creation of computational infrastructure to support effective stochastic computing. We view Monte Carlo algorithms as any algorithms that use random number to compute quantities of interest. As such, we have developed extremely efficient Monte Carlo for problems ranging from the computation of permeabilities of packed beds to the electrostatic field elicited on large biomolecules in ionic solution. In addition, we study basic computational kernels, such as those involved in numerical linear algebra, to find effective Monte Carlo alternatives to existing, deterministic, algorithms. Finally, we have been involved in the development of computational infrastructure for stochastic computing. We developed the Scalable Parallel Random Number Generators (SPRNG) library that is very widely used in highly demanding Monte Carlo applications. SPRNG was an outgrowth of our own research in pseudo- and quasirandom number generation which continues in concert with SPRNG development. Also, building on SPRNG, we have developed extensive infrastructure for stochastic computing on the computational grid. In all these activities we have had undergraduate and graduate students, as well as postdoctoral scholars, visitors, and international collaborators working with me in our research group.