当前位置: X-MOL 学术Comput. Appl. Math. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Efficiency improvement of Kriging surrogate model by subset simulation in implicit expression problems
Computational and Applied Mathematics ( IF 2.5 ) Pub Date : 2020-04-04 , DOI: 10.1007/s40314-020-01147-1
Liu Chu , Jiajia Shi , Eduardo Souza de Cursi , Shujun Ben

In practical engineering and industry fields, complicated and correlated problems are often descripted by implicit expression. The Kriging model is one of the popular spatial interpolation models to surrogate the numerical relationship between input and output variables. But the efficiency of the Kriging surrogate model is limited when confronting with large databases. The subset simulation is a promising selection method to provide more important and typical samples. By the subset simulation, the Kriging surrogate model can significantly reduce the computational cost in regression, since much fewer samples are required. Besides, more reliable prediction results can be obtained because of the emphasis on the samples that are more representative in the Kriging fitting process. Examples are performed to confirm the properties of the Kriging surrogate model based on the subset simulation.

中文翻译:

隐式表达式问题中子集模拟对Kriging代理模型效率的提高

在实际的工程和工业领域中,通常通过隐式表达来描述复杂且相关的问题。克里格模型是替代输入和输出变量之间的数值关系的一种流行的空间插值模型。但是,当面对大型数据库时,克里格代理模型的效率是有限的。子集模拟是一种有前途的选择方法,可以提供更重要和更典型的样本。通过子集模拟,克里格代理模型可以显着降低回归计算的成本,因为所需的样本要少得多。此外,由于重点在于在克里格拟合过程中更具代表性的样本,因此可以获得更可靠的预测结果。
更新日期:2020-04-04
down
wechat
bug