Quasi-Physical And Quasi-Human Algorithm For Forecasting Gold Atom Clusters Structure

A cluster is between micro-world and macro-world of the material world, has a unique physical and chemical property. Clusters nature of the structure is determined by its quality; therefore, to determine the structure of clusters has become an important issue which concern by the scientific community of cluster. The cluster structure is formed by the cluster composed of atoms and atomic number of the interaction potential decision. From the role of theory, atomic clusters potential function can be divided into three categories: experience in the situation, Based on first-principles self-consistent potential and the tight-binding potential. Clusters current forecast of the study were selected in three categories of potential function of a potential function of the circumstances, use of the potential function calculation the minimum of the potential energy clusters, to predict the structure of cluster of the minimum energy. Clusters structure prediction problem is a NP-Hard problem, the mainstream with the Monte Carlo algorithm, Simulated Annealing algorithm and Basin-Hopping methods.The gold clusters because of their own value and the nano-scale equipment and catalyst of potential uses has become increasingly significance. Thus, from a practical and easy analysis of the angle we choose Lennard-Jones potential energy function to predict their stable structure. Specific research methods use the Quasi-physical and Quasi-human of professor Huang.In this paper, the author firstly introduces the research methods and situation of the Au cluster structure problem. Then based on the review work of the computational complexity and NP-Hard problems, the author propose the general solving idea about NP hard problem and technical ways of corresponding quasi-human method to solve this problem. In order to facilitate the research, the author establishes a mathematical model for Au atomic cluster structure problem. And the main objective is heightened the utilization of the container space. On this basis, this paper researches the initial pattern generation algorithm and the gradient algorithm. According to the actual situation, improve the gradient algorithm. Finally the authors propose a new quasi-algorithm for Au atomic cluster problem based on professor Huang's years of study results and experience of human life and society. Through testing algorithm, the algorithm's actual results significantly, the number of atom between 2 and 51 examples have reached the international advanced level.