Dear @hokru , thanks for reply.
I’m trying to fit parameters of the non-bonded potential of force fields through the energy of dimer interaction. I know there are supermolecular and perturbative approaches (SAPT).
According to the literature (which usually considers the supermolecular approach), one of the ways to do this would be to calculate the energy gradient and obtain the forces experienced by each atom.
With the forces experienced by each atom by the quantum mechanics data, I would have to vary parameters of the non-bonded potential of the force field until I could reproduce the forces observed in quantum mechanics by molecular dynamics simulations. These are called force fields based on quantum mechanics (a bottom-up approach).
My idea would be to compare supermolecular and perturbative approaches (of course, if it makes sense) in obtaining parameters of the non-bonded potential of force fields. However, as it is apparently not possible to calculate forces by the perturbative approach, in this method of study the approaches would not be comparable I think now.