DLS Algorithm For Predicting The Ground State Structure Of Au Clusters

Cluster is a relatively stable aggregate which is composed of the binding force by physical or chemical. Different numbers of atoms make cluster have different physical and chemical property. The most important problem for cluster science is ensure the atoms and molecules how to evolve to the Cluster, and research the structure of cluster ant its property how to change by the evolved. The structure is the most basic characteristics of cluster, which directly affect other properties of clusters. So predict the structure become more important. As we all know, due to the increased number of atoms in the cluster make the model space increases exponentially, and this problem has been proved to be a NP-hard problem. People made a lot of heuristic algorithms to solve the problem, which is in polynomial time complexity can be solved exactly. Such as BH(Basin-Hopping) algorithm and its modification, GA(Genetic Algorithms), Simulated Annealing algorithm.The ground state structure prediction is a global optimization problem, This paper discusses the issue of potential energy function for cluster, the algorithm which can solve these problem, detailed description of a few of the classic heuristic algorithm, and base on Sutton-Chen Energy Model proposed DLS-IO algorithm. By the DLS-IO algorithm we make a lot of experiment, and the result show this algorithm can got some more outstanding cluster. For the Au3~80, we can get the current lowest energy structure. And at Au100, Au120, Au150, we can update the current lowest energy of the world.