Format

Send to

Choose Destination
Phys Chem Chem Phys. 2010 Oct 21;12(39):12450-6. doi: 10.1039/c0cp00373e. Epub 2010 Aug 18.

Computation of nodal surfaces in fixed-node diffusion Monte Carlo calculations using a genetic algorithm.

Author information

1
Department of Chemistry and Biochemistry, Utah State University, Logan, UT 84322-0300, USA. jordan.ramilowski@aggiemail.usu.edu

Abstract

The fixed-node diffusion Monte Carlo (DMC) algorithm is a powerful way of computing excited state energies in a remarkably diverse number of contexts in quantum chemistry and physics. The main difficulty in implementing the procedure lies in obtaining a good estimate of the nodal surface of the excited state in question. Although the nodal surface can sometimes be obtained from symmetry or by making approximations this is not always the case. In any event, nodal surfaces are usually obtained in an ad hoc way. In fact, the search for nodal surfaces can be formulated as an optimization problem within the DMC procedure itself. Here we investigate the use of a genetic algorithm to systematically and automatically compute nodal surfaces. Application is made to the computation of excited states of the HCN-(4)He complex and to the computation of tunneling splittings in the hydrogen bonded HCl-HCl complex.

PMID:
20717596
DOI:
10.1039/c0cp00373e

Supplemental Content

Full text links

Icon for Royal Society of Chemistry
Loading ...
Support Center