计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (24): 32-36.

• 学术探讨 • 上一篇    下一篇

基于最小生成树NSGA-2算法的改进

李密青,郑金华   

  1. 湘潭大学 信息工程学院,湖南 湘潭 411105
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-08-21 发布日期:2007-08-21
  • 通讯作者: 李密青

Improved NSGA-2 algorithm based on minimum spanning tree

LI Mi-qing,ZHENG Jin-hua   

  1. Institute of Information Engineering,Xiangtan University,Xiangtan,Hunan 411105,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-08-21 Published:2007-08-21
  • Contact: LI Mi-qing

摘要: 多目标进化算法(MOEA)的一个关键就是保持解的分布度,提出了一种用最小生成树的边的权值来表示个体聚集距离的方法,并且对NSGA-2的交叉算子和变异率进行了改进。实验结果表明,与NSGA-2相比该方法(MST-NSGA-2)在解的分布度上有较大的提高,并且有着良好的收敛性。

关键词: 多目标进化算法, 聚集距离, 最小生成树, 非均匀算术交叉

Abstract: The key of Multi-Objective Evolutionary Algorithm(MOEA) is how to keep diversity of solutions.In this paper,it suggests a new approach to measure individual crowding distance by edge weight of minimum spanning tree,and it is also improved crossover operator and mutation rate for NSGA-2.It’s shown by experimental results that the convergence and diversity of this algorithm is more efficient than NSGA-2.

Key words: Multi-Objective Evolutionary Algorithm(MOEA), crowding distance, minimum spanning tree, nonuniform arithmetical
crossover