A parallel evolutionary algorithm to search for global minima geometries of heterogeneous ab initio atomic clusters

In this work we present a parallel evolutionary algorithm applied to the search of the lowest-energy structures of heterogeneous atomic clusters. A new and improved crossover operator is proposed in order to always ensure the creation of new clusters with the same number of atomic elements. The approach proposed proved to be efficient and fast as all cluster calculations were performed by an ab initio quantum mechanics method, which is computationally expensive. Results of our search, obtained using the proposed approach, have been compared with previous calculations, and the efficiency has been confirmed, as we were able to find the global minimum and propose a wide number of new isomers of low-energy. Specifically, we addressed the problem to deal with clusters of lithium and fluorine atoms. However, the proposed algorithm can be extended to all kind of atomic and molecular clusters.