Genetic Algorithm for quick finding of diatomic molecule potential parameters

TL;DR

This paper demonstrates that a Genetic Algorithm can effectively determine parameters of diatomic molecule potentials, improving model accuracy over traditional Morse potentials using experimental spectra.

Contribution

The paper introduces a GA-based method for accurately fitting diatomic molecule potential parameters, including application to experimental data with improved models.

Findings

GA accurately reproduces potential parameters on synthetic data.

Switching from Morse to expanded Morse oscillator improves fit to experimental spectra.

GA-based fitting enhances the modeling of diatomic molecule potentials.

Abstract

Application of Genetic Algorithm (GA) for determination of parameters of an analytical representation of diatomic molecule potential is presented. GA can be used for finding potential characteristics of an electronic energy state which can be described by analytical function. GA was tested on two artificially generated datasets which base on potentials with known characteristics and two LIF excitation spectra recorded using transitions in CdKr and CdAr molecules. Tests on generated datasets showed that GA can properly reproduce parameters of the potentials. Tests on experimental spectra indicated that changing the potential model from Morse, which is frequently used as a starting potential in IPA, to expanded Morse oscillator (EMO) leads to noticeable improvement of agreement between simulated and experimental data.