Comparing Computer Experiments for Fitting High-Order Polynomial Metamodels

• Autores: Rachel T. Johnson, Douglas C. Montgomery, Bradley Jones, Peter A. Parker
• Localización: Journal of quality technology: A quarterly journal of methods applications and related topics, ISSN 0022-4065, Vol. 42, Nº. 1, 2010, págs. 86-102
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • When using computer simulation as a modeling tool, the underlying function of the computer experiment often results in too much curvature to be properly modeled by a low-order polynomial. Finding an appropriate design in such cases can be complicated. Several computer experiments are evaluated for the purpose of fitting a high-order polynomial to the response data from both optimal and space-filling designs. In addition, a class of hybrid designs is proposed that can be used for deterministic and stochastic simulation models.