Analysis of Multi-domain Complex Simulation Studies

  • James R. Gattiker
  • Earl Lawrence
  • David Higdon
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3982)


Complex simulations are increasingly important in systems analysis and design. In some cases simulations can be exhaustively validated against experiment and taken to be implicitly accurate. However, in domains where only limited validation of the simulations can be performed, implications of simulation studies have historically been qualitative. Validation is notably difficult in cases where experiments are expensive or otherwise prohibitive, where experimental effects are difficult to measure, and where models are thought to have unaccounted systematic error. This paper describes an approach to integrate simulation experiments with empirical data that has been applied successfully in a number of domains. This methodology generates coherent estimates of confidence in model predictions, model parameters, and estimates, i.e. calibrations, for unobserved variables. Extensions are described to integrate the results of separate experiments into a single estimate for simulation parameters, which demonstrates a new approach to model-based data fusion.


Simulation Response Gravitational Attraction Thermal Activation Energy Variable Importance Measure Maximum Yield Stress 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Los Alamos Science special issue on Science-Based Prediction for Complex Systems  29 (2005)Google Scholar
  2. 2.
    Gattiker, J.: Using the Gaussian Process Model for Simulation Analysis Code. Los Alamos technical report LA-UR-05-5215 (2005)Google Scholar
  3. 3.
    Christie, G., Grove, H., Sharp, S.: Error Analysis in Simulation of Complex Phenomena. Los Alamos Science special issue on Science-Based Prediction for Complex Systems 29, 6–25 (2005)Google Scholar
  4. 4.
    Chib, S., Greenberg, E.: Understanding the Metropolis-Hastings Algorithm. The American Statistician 49(4), 327 (1995)CrossRefGoogle Scholar
  5. 5.
    Fugate, M., Williams, B., Higdon, D., Hanson, K., Gattiker, J., Chen, S., Unal, C.: Hierarchical Bayesian Analysis and the Preston-Tonks-Wallace Model, Los Alamos Technical Report, LA-UR-05-3935 (2005)Google Scholar
  6. 6.
    Kennedy, M., O’Hagan, A.: Bayesian Calibration of Computer Models (with discussion). Journal of the Royal Statistical Society B 68, 426–464 Google Scholar
  7. 7.
    Jones, D., Schonlau, M., Welch, W.: Efficient Global Optimization of Expensive Black-Box Functions. Journal of Global Optimization 13, 455–492 (1998)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Higdon, D., Kennedy, M., Cavendish, J., Cafeo, J., Ryne, R.: Combining field Observations and Simulations for Calibration and Prediction. SIAM Journal of Scientific Computing 26, 448–466 Google Scholar
  9. 9.
    Nakhleh, C., Higdon, D., Allen, C., Kumar, V.: Bayesian Reconstruction of Particle Beam Phase Space from Low Dimensional Data, Los Alamos technical report LA-UR-05-5897Google Scholar
  10. 10.
    Higdon, D., Gattiker, J., Williams, B., Rightley, M.: Computer Model Calibration using High Dimensional Output, Los Alamos Technical report LAUR-05-6410, submitted to the Journal of the American Statistical AssociationGoogle Scholar
  11. 11.
    Williams, B., Higdon, D., Gattiker, J.: Uncertainty Quantification for Combining Experimental Data and Computer Simulations, Los Alamos Technical Report LA-UR-05-7812Google Scholar
  12. 12.
    Jones, D.R., Schonlau, M., Welch, W.: Efficient Global Optimization of Expensive Balck-Box Functions. Journal of Global Optimization 13, 455–492 (1998)MATHCrossRefMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • James R. Gattiker
    • 1
  • Earl Lawrence
    • 1
  • David Higdon
    • 1
  1. 1.Los Alamos National Laboratory 

Personalised recommendations