Guiding-function optimization in biased-selection Monte Carlo calculations
Abstract
We show that the guiding function, which is used to generate the Monte Carlo integration points in a biased-selection calculation, can be optimized so as to minimize the statistical error of an expectation value. For excited states we introduce an orthogonalizing part into the minimization functional to prevent ground and excited states from mixing.
- Publication:
-
Journal of Chemical Physics
- Pub Date:
- December 1992
- DOI:
- 10.1063/1.463410
- Bibcode:
- 1992JChPh..97.8407A