Optimization of a molecular mechanics force field for polyoxometalates based on a genetic algorithm.

A stochastic technique based on genetic algorithms was implemented to develop new force fields by optimizing molecular mechanics (MM) parameters. These force fields have been optimized for inorganic compounds such as polyoxometalates (POMs) and especially for type-I polymolybdate and polytungstate clusters. Focussing on the methodology of the development of the force fields, they were tested for the prediction of structural parameters, comparing the MM optimized structures with the geometry obtained after an optimization based on density functional theory. Results show that the genetic algorithm converges toward an optimum combination of parameters which successfully reproduces POMs structures with a high degree of accuracy.