Multiscale model modeling

CMA 3

Christopher P. Calderon, calderoc224@hotmail.com, Department of Statistics and Department of Computational and Applied Mathematics, Rice University, P O Box 1892 (MS#138), Houston, TX 77251-1892
Methods for computing equilibrium properties from nonequilibrium simulations/data have received a lot of attention in biophysics due to their potential relevance to a wide variety of systems. I briefly discuss an established result (the Jarzynski Equality) which can be used to develop sampling schemes that utilize nonequilibrium control strategies to force (and determine the energy change associated with) rapid state changes in molecular systems in a preset amount of simulation time. Obtaining high accuracy within this type of framework often requires a large number of independent nonequilibrium trajectories. The number of realizations needed for acceptable accuracy is typically too large for practical computations; this motivates the surrogate process approximations (SPAs) introduced and discussed in this talk. Various intriguing features of this approximation are discussed, e.g. a collection of different (but related) SPAs are needed to accurately describe the full detailed system dynamics. I also discuss how this work would significantly benefit from an interdisciplinary collaboration of researchers.