This presentation will be given online by zoom, with the zoom link sent to the AMC seminar list. Participants interested in viewing should sign up for that list (see AMC Seminar website).
ABSTRACT: The efficient solution of systems of nonlinear equations is an important tool for the modeling of physical phenomena. We will disucss the use and theoretical analysis of an extrapolation method known as Anderson acceleration to improve the convergence properties of fixed-point methods to solve nonlinear systems. We will further look at how to improve the performance of the method by adaptively updating the extrapolation parameters; and, consider some extensions not yet covered by the theory.
BIO: Sara Pollock is an Assistant Professor at University of Florida. She earned her Ph.D. in Mathematics with a specialization in Computational Science from the University of California, San Diego, in 2012. She also holds an M.S. in Applied Mathematics from the University of Washington, Seattle, a B.S. in Mathematics from the University of New Mexico, and a Studio Art Diploma from the School of the Museum of Fine Arts, Boston. She works on finite element methods for nonlinear and multiscale problems, and developing efficient and robust solution techniques together with well-posedness results for discrete nonlinear systems.