Premium
Fitting complex potential energy surfaces to simple model potentials: Application of the simplex‐annealing method
Author(s) -
Bustos Marún Raúl A.,
Coronado Eduardo A.,
Ferrero Juan C.
Publication year - 2005
Publication title -
journal of computational chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.907
H-Index - 188
eISSN - 1096-987X
pISSN - 0192-8651
DOI - 10.1002/jcc.20168
Subject(s) - simulated annealing , potential energy , parameterized complexity , simplex algorithm , simplex , potential energy surface , mathematics , simple (philosophy) , robustness (evolution) , pair potential , statistical physics , mathematical optimization , algorithm , chemistry , physics , thermodynamics , linear programming , geometry , quantum mechanics , philosophy , biochemistry , epistemology , ab initio , gene
A stochastic method of optimization, which combines simulated annealing with simplex, is implemented to fit the parameters of a simple model potential. The main characteristic of the method is that it explores the whole space of the parameters of the model potential, and therefore it is very efficient in locating the global minimum of the cost function, in addition to being independent of the initial guess of the parameters. The method is employed to fit the complex intermolecular potential energy surface of the dimer of water, using as a reference the spectroscopic quality anisotropic site–site potential of Feller et al. The simple model potential chosen for its reparameterization is the MCY model potential of Clementi et al. The quality of the fit is assessed by comparing the geometry of the minimum, the harmonic frequencies, and the second virial coefficients of the parameterized potential with the reference one. Finally, to prove more rigorously the robustness of this method, it is compared with standard nonstochastic methods of optimization. © 2005 Wiley Periodicals, Inc. J Comput Chem 26: 523–531, 2005