西安电子科技大学学报 ›› 2016, Vol. 43 ›› Issue (2): 157-161+192.doi: 10.3969/j.issn.1001-2400.2016.02.027

• 研究论文 • 上一篇    下一篇

求解多目标考试时间表问题的NNIA改进算法

雷雨;焦李成;公茂果;李玲玲   

  1. (西安电子科技大学 智能感知与图像理解教育部重点实验室,陕西 西安  710071)
  • 收稿日期:2015-01-08 出版日期:2016-04-20 发布日期:2016-05-27
  • 通讯作者: 雷雨
  • 作者简介:雷雨(1985-), 男,西安电子科技大学博士研究生,E-mail:xdleiyu@gmail.com.
  • 基金资助:

    国家自然科学基金资助项目(61273317)

Enhanced NNIA for multi-objective examination timetabling problems

LEI Yu;JIAO Licheng;GONG Maoguo;LI Lingling   

  1. (Ministry of Education Key Lab. of Intelligent Perception and Image Understanding, Xidian Univ., Xi'an  710071, China)
  • Received:2015-01-08 Online:2016-04-20 Published:2016-05-27
  • Contact: LEI Yu

摘要:

针对多目标考试时间表问题,提出了一种用于求解多目标考试时间表问题的非支配邻域免疫改进算法.采用经典多目标进化算法非支配邻域免疫的算法框架,为改善该算法的收敛性和多样性,利用超启发方法生成初始种群,同时采用一种新的资源分配模型,在克隆过程中动态调节优秀个体的克隆比例,来改善算法的性能.采用10组标准测试数据对算法的性能进行测试.实验结果表明,该算法对多目标考试时间表问题十分有效,并且在某些测试数据上可得到令人满意的结果.

关键词: 多目标优化, 考试时间表, 进化算法, 资源分配模型, 非支配邻域免疫算法

Abstract:

Based on the nondominated neighbor immune algorithm(NNIA), an enhanced NNIA is introduced for multi-objective examination timetabling problems. With the framework of NNIA, the hyper-heuristic approach is utilized to generate the initial population. In addition, the resource allocation model is designed to dynamically adjust the clone percentage of potential individuals. Experimental results on ten benchmark datasets prove that the proposed algorithm can solve examination timetabling problems effectively and obtaine competitive results.

Key words: multi-objective optimization, examination timetabling, evolutionary algorithm, resource allocation model, nondominated neighbor immune algorithm(NNIA)

Baidu
map