Automatic construction of <em>ab initio</em> potential energy surfaces

Monday, January 12, 2009 - 11:10am - 11:40am
EE/CS 3-180
Donald Thompson (University of Missouri)
A highly accurate and efficient method for molecular global
potential energy surface (PES) construction and fitting is
demonstrated. An interpolating moving least-squares (IMLS)
method using low-density ab initio potential, gradient, or
Hessian values to compute PES parameters is shown to lead to an
accurate and efficient PES representation. The method is
automated and flexible so that a PES can be optimally generated
for classical trajectories, spectroscopy, or other
applications. Two main drivers for the fitting method have
been developed thus far. The first is a PES generator designed
primarily for spectroscopy applications. Using this method,
the configuration space defined by a specified energy range is
automatically fit to a predefined accuracy. A second approach
is based on trajectory methods for computing reaction rates.
In this approach, the configuration space that is dynamically
accessible to a particular ensemble of trajectories is fit on
the fly. Results that are indicative of the accuracy,
efficiency, and scalability will be presented.

MSC Code: