Differential evolution algorithm approach for describing vibrational solvatochromism

We model the solvation-induced vibrational frequency shifts of the amide I and amide II modes of N-methylacetamide in water and the nitrile stretch mode of acetonitrile in water by expressing the frequency shift as a polynomial function expanded by the inverse power of interatomic distances. The coefficients of the polynomial are optimized to minimize the deviation between the predicted frequency shifts and those calculated with quantum chemistry methods. Here, we show that a differential evolution algorithm combined with singular value decomposition is useful to find the optimum set of coefficients of polynomial terms. The differential evolution optimization shows that only a few terms in the polynomial are dominant in the contribution to the vibrational frequency shifts. We anticipate that the present work paves the way for further developing different genetic algorithms and machine learning schemes for their applications to vibrational spectroscopic studies.We model the solvation-induced vibrational frequency shifts of the amide I and amide II modes of N-methylacetamide in water and the nitrile stretch mode of acetonitrile in water by expressing the frequency shift as a polynomial function expanded by the inverse power of interatomic distances. The coefficients of the polynomial are optimized to minimize the deviation between the predicted frequency shifts and those calculated with quantum chemistry methods. Here, we show that a differential evolution algorithm combined with singular value decomposition is useful to find the optimum set of coefficients of polynomial terms. The differential evolution optimization shows that only a few terms in the polynomial are dominant in the contribution to the vibrational frequency shifts. We anticipate that the present work paves the way for further developing different genetic algorithms and machine learning schemes for their applications to vibrational spectroscopic studies.