中国科学技术大学学报 ›› 2015, Vol. 45 ›› Issue (5): 373-378.DOI: 10.3969/j.issn.0253-2778.2015.05.005

• 原创论文 • 上一篇    下一篇

基于盆地跳和极小值跳算法的兰纳-琼斯簇 全局优化对比研究

姚 成   

  1. 1.安徽财贸职业学院电子信息系,安徽合肥 230601; 2.Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA; 3.中科院大连化学物理研究所, 辽宁大连 116000;4.安徽大学计算智能与信号处理教育部重点实验室,安徽合肥 230039
  • 收稿日期:2015-03-24 修回日期:2015-05-12 出版日期:2015-05-31 发布日期:2015-05-31
  • 通讯作者: 姚成
  • 作者简介:姚成(通讯作者),男,1964年生,副教授. 研究方向:算法研究. E-mail: yao234@163.com
  • 基金资助:
    安徽省高等教育振兴计划专业带头人培养资助项目(皖教秘人〔2013〕189号),安徽省自然科学基金(1408085QF126)资助.

Comparative study of basin-hopping and the minima-hopping global optimization methods: A case study on Lennard-Jones clusters

YAO Cheng   

  1. 1.Department of Electronic Information, Anhui Finance & Trade Vocational College, Hefei 230601, China; 2.Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA; 3.Dalian Institute of Chemical Physics of Chinese Academy of Sciences, Dalian 116000, China; 4.Key Laboratory of Intelligent Computing & Signal Processing, Ministry of Education, Anhui University, Hefei 230039, China
  • Received:2015-03-24 Revised:2015-05-12 Online:2015-05-31 Published:2015-05-31

摘要: 为探索寻找经典模型兰纳-琼斯簇的全局最优构型的更优算法,选取盆地跳和极小值跳算法,计算了13粒子、38粒子以及60粒子的兰纳-琼斯簇最优构型.结果表明:这两种算法都可以成功找到最优构型,其中极小值跳方法相比较而言有较快的搜索速度,它的搜索步数约比盆地跳算法少至少一个数量级.

关键词: 全局最优化, 盆地跳算法, 极小值跳算法, 兰纳-琼斯簇

Abstract: In order to find a better algorithm for global optimization of Lennard-Jones cluster, two algorithms, basin-hopping and minima-hopping, were used to optimize the structures of 13 particles, 38 particles, and 60 particles Lennard-Jones clusters. Both algorithms find global minima, while the minima-hopping method yields a faster search speed which is at least one order faster than the basin-hopping method.

Key words: global optimization, basin-hopping method, minima-hopping method, Lennard-Jones cluster