引用本文: | 刘新亮,郭波.适用于复杂系统仿真试验的试验设计方法[J].国防科技大学学报,2009,31(6):95-99.[点击复制] |
LIU Xinliang,GUO Bo.Studies on the Optimization of Latin hypercube designs for Complex System Simulation[J].Journal of National University of Defense Technology,2009,31(6):95-99[点击复制] |
|
|
|
本文已被:浏览 6453次 下载 5729次 |
适用于复杂系统仿真试验的试验设计方法 |
刘新亮, 郭波 |
(国防科技大学 信息系统与管理学院,湖南 长沙 410073)
|
摘要: |
拉丁超立方试验设计是“充满空间”试验设计方法的重要领域。传统拉丁超立方试验设计优化以列相关系数最小化或试验点之间最大距离最小为单一优化准则,这种单目标试验设计优化方法存在缺陷。提出了将两类优化准则作为优化目标的多目标优化准则,以及实现多目标优化的改进ESE算法。算例分析证明,提出的试验设计优化算法优于已有典型试验设计方法。 |
关键词: 复杂系统仿真试验 拉丁超立方设计 多目标优化 |
DOI: |
投稿日期:2009-04-08 |
基金项目:国家部委资助项目 |
|
Studies on the Optimization of Latin hypercube designs for Complex System Simulation |
LIU Xinliang, GUO Bo |
(College of Information System and Management, National Univ. of Defense Technology, Changsha 410073, China)
|
Abstract: |
The optimization of Latin hypercube design is an important area of the space filling experimentations. The existing procedures to find optimal LHD by minimizing the pairwise correlations or maximizing the minimum inter-site distance in traditional LHD optimization are introduced, and their drawbacks are discussed in these methods. This paper proposes a multi-objective optimization approach to find optimal LHD by combining pairwise correlation and the maximin inter-site distance performance measures by modified ESE algorithm. Several examples are presented to show that the optimal designs are good in terms of both the correlation and distance criteria. |
Keywords: complex system simulation Latin hypercube designs multi-objective optimization |
|
|