Numerical Methods - Part 2

Michael Holst
University of California, San Diego
Mathematics

In the second lecture, we will focus on an application arising in biophysics, namely the nonlinear Poisson-Boltzmann equation model
of the electrostatics of biomolecules. We will first give an overview of the solution theory for the equation, followed by a priori and
a posteriori error estimates for finite element methods. We then describe an unusual parallel algorithm for use with the adaptive
algorithm described in the first lecture, and derive some local and global error estimates for the behavior of the parallel algorithm.
We then apply the algorithm to the nonlinear Poisson-Boltzmann equation on large Beowulf-class parallel clusters.


Back to Multiscale Modeling Tutorials