INDEPENDENT VARIATION OF THE DENSITY AND THE POTENTIAL IN DENSITY-FUNCTIONAL METHODS

It is possible to write down an energy functional within density-functional theory that is maximal respective to the potential and minimal respective to the density when these are varied as independent quantities. Based on this property, methods can be devised that directly optimize the density and the potential respective to some chosen degrees of freedom. The main advantage of such an approach compared to full-potential methods is that the major calculational steps must only be done for functions of a selected simpler form. Test calculations for the N-2 dimer show that optimization in a very-low-dimensional space can already give good accuracy.