Bkapproximation applied to CI-SDTQ
Vladimı́r Sychrovský, Petr Čárský · Molecular Physics · 1996
We have applied the B k approximation to single reference CI-SDTQ with the aim of examining whether this CI variant may be useful for practical calculations in terms of its accuracy and cost. The few test calculations we have performed suggest that the energy error relative to the CI-SDTQ energy is reduced on average by 80% when passing from CI-SD to CI-SDTQ-B k . Also the size-inconsistency error is reduced. In our test calculations on (H2O)2, the CI-SD error was 0·00475 au and the CI-SDTQ-B k error was 0·00043 au, close to the error of 0·00034 au given by CI-SDTQ. From the number of processed loops in GUGA treatments, it may be concluded that with an efficient coded computer program the cost of CI-SDTQ-B k should be moderately higher than that of CI-SD, but considerably lower than the cost of CI-SDTQ.