A general construction for nested Latin hypercube designs

被引：9
|
作者：
Xu, Jin [1 ]
Duan, Xiaojun [1 ]
Wang, Zhengming [1 ]
Yan, Liang [1 ]
机构：
[1] Natl Univ Def Technol, Coll Sci, Changsha 410073, Hunan, Peoples R China
来源：
STATISTICS & PROBABILITY LETTERS | 2018年 / 134卷
基金：
中国国家自然科学基金;
关键词：
Computer experiments; Latin hypercube designs; Nested designs; Sampling property; COMPUTER EXPERIMENTS; HIGH-ACCURACY; POINT;
D O I：
10.1016/j.spl.2017.10.022
中图分类号：
O21 [概率论与数理统计]; C8 [统计学];
学科分类号：
020208 ; 070103 ; 0714 ;
摘要：
We propose a new construction for nested designs, called General Nested Latin Hypercube designs (GNLHs). Such designs contain nested Latin hypercube designs as special cases. Besides achieving maximum uniformity in one dimension, each layer of GNLHs is flexible in run sizes. Moreover, theoretical results and numerical simulations show that GNLHs perform well on the sampling variance. (C) 2017 Elsevier B.V. All rights reserved.
引用
收藏
页码：134 / 140
页数：7
相关论文
共 50 条
12345