Eigenvalues calculations with Gaussian functions

If one does not directly fit the numerical potential to a Gaussian expansion, the exact analytical eigenfunctions and eigenvalues (mentioned previously) are used to generate analytical forms of the potentials by the method proposed by Barthelat and Durand. As already stated, there is a need to balance computational effort with desired accuracy. The more compact the potentials are made, the greater the computational savings when the ECPs are used in calculations. On the other hand, agreement between the exact and optimized values should be as close as possible. The optimization process is achieved by minimization of the functional in Eq. [5], where quantities capped... 

---