A Hierarchical Metamodeling Method for Large Scale, Multi-Objective Computer Simulations

This paper introduces a new hierarchical metamodeling scheme for large scale computer simulations. The goal of the this new metamodeling method is to allow for the creation of metamodels that are as accurate as traditional metamodeling schemes but can be created with considerably fewer data points. In addition the new method allows the model builder to create hierarchically partitioned metamodels of black box systems such as finite element models. Black box is this case refers to systems that do not have intermediate variables/ responses around which to partition the analysis. The hierarchical metamodeling method is demonstrated on a finite element model consisting of 40 input variables and 32 responses.

[1]  Paul K. Davis,et al.  Preface , 2021 .

[2]  Don R. Holcomb,et al.  Some notes on the construction and evaluation of supersaturated designs , 2002 .

[3]  Dimitri N. Mavris,et al.  Assessing the Simultaneous Impacts of Requirements, Vehicle Characteristics, and Technologies During Aircraft Design , 2001 .

[4]  Dimitri N. Mavris,et al.  Preliminary Evaluation of a Hierarchical Metamodeling Technique for Systems with a Large Number of Parameters , 2004 .

[5]  Eric Fox,et al.  The Pratt & Whitney Probabilistic Design System , 1994 .

[6]  Kenneth H. Rosen,et al.  Discrete Mathematics and its applications , 2000 .

[7]  R. H. Myers,et al.  Response Surface Methodology: Process and Product Optimization Using Designed Experiments , 1995 .

[8]  Dimitri N. Mavris,et al.  Assessment of the Efficiency of Partitioned Response Surface Equations in Structural Applications , 2003 .

[9]  A. Owen.  Controlling correlations in latin hypercube samples , 1994 .

[10]  Thomas D. Wickens,et al.  The geometry of multivariate statistics , 1994 .

[11]  Yan-Gang Zhao,et al.  A general procedure for first/second-order reliabilitymethod (FORM/SORM) , 1999 .

[12]  Dimitri N. Mavris,et al.  Overview of Metamodeling Techniques for Problems with a Large Number of Input Parameters , 2003 .

[13]  Eric Fox,et al.  On the accuracy of various probabilistic methods , 2000 .

[14]  Dimitri N. Mavris,et al.  Partitioned, multilevel response surfaces for modeling complex systems , 2000 .

[15]  Thomas J. Santner,et al.  The Design and Analysis of Computer Experiments , 2003, Springer Series in Statistics.

[16]  John Marc Zentner,et al.  A Design Space Exploration Process for Large Scale, Multi-Objective Computer Simulations , 2006 .

[17]  Vicente J. Romero,et al.  Toward a New Class of Hybridized Reliability Methods designed for Efficiency and Robustness with Large Computer Models , 2001 .

[18]  Andy J. Keane,et al.  Engineering Design via Surrogate Modelling - A Practical Guide , 2008 .