宇航学报 ›› 2012, Vol. 33 ›› Issue (2): 249-253.doi: 10.3873/j.issn.1000-1328.2012.02.015

• 电子信息 • 上一篇    下一篇

基于遗传算法的重定位容错方法的研究

仉俊峰1,2, 陈德运1, 洪炳镕3, 苏健民2   

  1. 1. 哈尔滨理工大学仪器科学与技术博士后流动站, 哈尔滨 150080;
    2. 东北林业大学,哈尔滨 150040;
    3. 哈尔滨工业大学, 哈尔滨 150001
  • 收稿日期:2011-02-24 修回日期:2011-05-17 出版日期:2012-02-15 发布日期:2012-02-21
  • 作者简介:1000-1328(2012)02-0249-05
  • 基金资助:
    中国博士后科学基金(20080440901);黑龙江省青年资金(QC08C29)

Research on Genetic-Based Algorithm Relocation Fault Tolerance Method

ZHANG Jun-feng1,2, CHEN De-yun1, HONG Bing-rong3, SU Jian-min2   

  1. 1.Instrument Science and Technology Postdoctoral Workstation, Harbin University of Science and Technology, Harbin 150080,China;
    2.Northeast Forestry University, Harbin 150040, China; 3.Harbin Institute of Technology, Harbin 150001, China
  • Received:2011-02-24 Revised:2011-05-17 Online:2012-02-15 Published:2012-02-21

摘要: 随着网络成为人们获取信息的主要工具,人们对网络可靠性的要求也越来越高。针对网络中的结点出现故障而影响网络传输效率和可靠性问题,进行了基于遗传算法的重定位容错方法的研究。首先,描述了树型网络结构,根据网络组件的开销提出了网络代价目标函数。其次,阐述了遗传算法最小代价路径的搜索过程,提出了基于遗传算法的重定位容错方法。最后,通过仿真实验验证了重定位容错方法在路径搜索中的有效性。这样既缩短了路径搜索时间,又提高了网络传输信息的可靠性。

关键词: 拓扑结构, 遗传算法, 最小代价路径, 重定位容错

Abstract: The network is becoming main tool for people fetching information, and people require high dependable network now. Because there are some faults in network nodes, which will effect on the efficiency and dependability of transmission. Genetic algorithm relocation fault tolerance method is studied in this paper. At first, the tree topological structures of networks is described, and the network costs-based objective function is put forward according to component costs of network. Second, the search process of genetic algorithm for minimum cost path is described. Meanwhile , the relocation fault tolerance method based on genetic algorithm is posed . At last, the method is verified by simulation experiments. Both the path search time is shorten and the dependability of network is improved in this way.

Key words: TopoloGical structure, genetic algorithm, Minimum cost path, Relocation fault tolerance