Dissertation
Dissertation > Mathematical sciences and chemical > Physics > Molecular physics, atomic physics > Molecular Physics

High-efficient Heuristic Algorithms on Prediction Ground State Structures of Clusters

Author NiHaiWen
Tutor XuRuChu;HuangWenQi
School Huazhong University of Science and Technology
Course Computer Software and Theory
Keywords Global optimization cluster optimization dynamic grids two-phase decline
CLC O561
Type Master's thesis
Year 2011
Downloads 18
Quotes 0
Download Dissertation

The prediction for the ground state structures of atomic clusters is one of the core issues of cluster research, which directly affect discussion of cluster. Now the simplification of potential energy function model is used to describe clusters, and by solving the lowest minimum of the potential energy function to get ground state structure of clusters. Although the potential function model is a simplification of the true model, but the problem is still NP-hard. Therefore, for clusters ground state structure prediction problem, the focus of current research is to design heuristic algorithms to solve the problem.The rare gas clusters, Au cluster and Co cluster was studied. By solving the potential energy function to determine the ground state structure of clusters, the main works contained in the thesis include: Firstly, the nature of the potential energy functions and how to generate initial structure of clusters were discussed. Secondly, Quasi-physical and Quasi- sociological algorithm was proposed to solve optimal structure of rare gas cluster.Thirdly, the TP-DLS algorithm which base on lattice- search and two-phase decline algorithm was designed to solve the structure of gas clusters, Au clusters and Co clusters. Some of the parameters in the potential energy function does not affect the results of the problem were found, which can be removed to simplify the model. For rare gas clusters, we solved its potential energy function by Quasi-physical and Quasi-sociological algorithm when atomic number between 38 and 74. All the results reached current optimal value.TP-DLS is not only an efficient algorithm, but also get new configurations of Au58 and Co101. Two new configurations of Au58 were obtained when we optimized Au clusters. The two new configurations are decahedron ,the energy of the two configurations are -15648.5689 and -15648.8754, they are all better than the current best energy value -15647.9269. The new configuration of Co101 was obtained when optimized Co cluster, the new configuration is also decahedron. The energy value of this new configuration is -384.329400, which is better than the current lowest energy value -384.1472457.

Related Dissertations
More Dissertations