Ramblings around CEPA-0

Ágnes Szabados
Laboratory of Theoretical Chemistry, Eötvös Loránd University, Budapest, Hongrie.

Jeudi 12 Décembre 2013, 14^h00
bibliothèque LCT, tour 12 - 13, 4e étage

The Coupled Electron Pair Approximation (CEPA) remains a valuable tool of electronic structure description even with the success of the more sophisticated Coupled Cluster (CC) theory. Methods of the CEPA family offer a cheaper alternative to CC and may be easier to adapt for the multireference case. The simplest variant among CEPA based methods, CEPA-0 is known under many names, for it is easy to arrive to it from various approaches. In many-body perturbation theory (MBPT) it was baptized D-MBPT(∞) in CC it can be recovered as linearized coupled cluster doubles, hence the designation LCCD.
The presentation will focus on facts less known about CEPA-0. We will work out its relationship with partitioning (i.e. level-shift) optimization in MBPT. This will take us to find a connection between CEPA-0 and the spin component scaled MP method, proposed by Grimme. Finally we will see yet another way to arrive to the CEPA-0 energy formula, based on Löwdin's bracketing function.