Dresden 2017 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 27: Computational Materials Modelling - Electronic structure approaches
MM 27.1: Vortrag
Dienstag, 21. März 2017, 10:15–10:30, IFW B
A Compressed-Sensing Approach to Select Accurate Atom-Centered Basis Functions for Advanced Density Functionals and Quantum Chemistry — •Niklas Menzel1, Luca M. Ghiringhelli1, Gitta Kutyniok2, and Matthias Scheffler1,3 — 1Fritz-Haber-Institut der MPG, Berlin, DE — 2Technische Universität, Berlin, DE — 3UC Santa Barbara, USA
The choice of basis sets is a crucial factor in density functional theory and quantum chemistry, determining the accuracy and computational cost of the calculation. Commonly used basis sets are not sufficiently accurate to represent the eigenfunctions for advanced exchange-correlation treatments. This leads to basis set extensions, such as the correlation-consistent basis sets by Dunning [JCP 90, 1007 (1989)]. We have developed a new approach to select basis functions based on compressed sensing (CS), a recently developed signal processing technique. CS provides a simple and efficient framework for basis set selection based on ℓ1-norm regularization. Our CS-based approach is applicable to all atom-centered basis sets. As introductory example, we demonstrate the selection of Gaussian-type basis functions (GTO). The number of chosen contracted or uncontracted GTOs can be tuned. We calculate the total energy for atoms from H to Ar and compare the results with Dunning’s correlation-consistent basis sets, which give similar results. Our new approach enables us to determine accurate basis sets for heavier atoms including d- and f-elements.