Enhanced procedure for the refinement of kinetic and potential functions for large-amplitude vibrations

Abstract

An enhanced method for the refinement of kinetic and potential terms for large-amplitude vibrations has been shown. To ensure the search for a minimum, the quadratic error between the observed and calculated energy levels is minimized with a quasi-Newton algorithm using a positive-definite Hessian matrix. In addition, analytical derivatives are obtained for computation of the gradient vector.