...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Quality Control and Applied Statistics >Replication or exploration? Sequential design for stochastic simulation experiments
【24h】

Replication or exploration? Sequential design for stochastic simulation experiments

机译：Replication or exploration? Sequential design for stochastic simulation experiments

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相关主题

摘要

Design and analysis of computer experiments use deterministic solvers from the physical sciences via Gaussian process (GP) interpolation. This type of modeling and stochastic simulation is common in social, management and biological sciences. The higher the noise, the more sophisticated the experiment required for tracking changes in noise. Replication is used to derive an estimate of the noise, though the primary interest is in the signal. The article focuses on a global surrogate model prediction of accuracy and computational efficiency. Wang and Haaland (Ref. 1) have shown that replicates can characterize the source of inaccuracy. Replication is useful in surrogate-assisted optimization of noisy blackbox functions. However, existing studies have not dealt with the decision problem for design of GP surrogates in the face of noisy simulations. In other words, it is important to know the allocation of unique locations and degree of replication. The article tackles the issue sequentially, with one design at a time, to find whether an element needs replication or a new location using integrated mean-square prediction error.

著录项

获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号