MAP: an MP2 accuracy predictor for weak interactions from adiabatic connection theory (bibtex)
by Stefan Vuckovic, Eduardo Fabiano, Paola Gori-Giorgi, Kieron Burke
Abstract:
Second order Møller-Plesset perturbation theory (MP2) approximates the exact Hartree-Fock (HF) adiabatic connection (AC) curve by a straight line. Thus by using the deviation of the exact curve from the linear behaviour, we construct an indicator for the accuracy of MP2. We then use an interpolation along the HF AC to transform the exact form of our indicator into a highly practical MP2 accuracy predictor (MAP) that comes at negligible additional computational cost. We show that this indicator is already applicable to systems that dissociate into fragments with a non-degenerate ground state, and we illustrate its usefulness by applying it to the S22 and S66 datasets
Reference:
MAP: an MP2 accuracy predictor for weak interactions from adiabatic connection theory Stefan Vuckovic, Eduardo Fabiano, Paola Gori-Giorgi, Kieron Burke, Submitted (2020).
Bibtex Entry:
@article{VFGB20,
	Pub-num 	   = {197},
	Title 		   = {MAP: an MP2 accuracy predictor for weak interactions from adiabatic connection theory},
	Author 		   = {Stefan Vuckovic, Eduardo Fabiano, Paola Gori-Giorgi, Kieron Burke},
	Abstract 	   = {Second order Møller-Plesset perturbation theory (MP2) approximates the exact Hartree-Fock (HF) adiabatic connection (AC) curve by a straight line. Thus by using the deviation of the exact curve from the linear behaviour, we construct an indicator for the accuracy of MP2. We then use an interpolation along the HF AC to transform the exact form of our indicator into a highly practical MP2 accuracy predictor (MAP) that comes at negligible additional computational cost. We show that this indicator is already applicable to systems that dissociate into fragments with a non-degenerate ground state, and we illustrate its usefulness by applying it to the S22 and S66 datasets},
%%	Doi 		   = {},
%%	Issn		   = {},
	Year 		   = {2020},
	Month 		   = {January},
	Journal		   = {Submitted},
%%	Volume 		   = {},
%%	Issue 		   = {},
%%	Number 		   = {},
%%	Pages 		   = {},
%%	Publisher 	   = {},
%%	Url 		   = {},
	arXiv		   = {2001.06364},
%%	keywords 	   = {}
%%}
Powered by bibtexbrowser