Uncertainty Analysis for Computationally Expensive Models with Multiple Outputs

详细信息    查看全文

• 作者：David Ruppert (1)
  Christine A. Shoemaker (2)
  Yilun Wang (3) (4)
  Yingxing Li (5)
  Nikolay Bliznyuk (6)
  
• 关键词：Bayesian calibration ; Computer experiments ; Groundwater modeling ; Inverse problems ; Markov chain Monte Carlo ; Radial basis functions ; SOARS ; Surrogate model ; SWAT model ; Town Brook watershed ; Uncertainty analysis
• 刊名：Journal of Agricultural, Biological, and Environmental Statistics
• 出版年：2012
• 出版时间：December 2012
• 年：2012
• 卷：17
• 期：4
• 页码：623-640
• 全文大小：611KB
• 参考文献：1. Arnold, J. G., Srinivasan, R., Muttiah, R. R., and Williams, J. R. (1998), 鈥淟arge Area Hydrologic Modeling and Assessment. Part I: Model Development,鈥? / Journal of the American Water Resources Association, 34, 73鈥?9. CrossRef
  2. Bates, D. M., and Watts, D. G. (1988), / Nonlinear Regression Analysis and its Applications, New York: Wiley.
  3. Bayarri, M. J., Berger, J. O., Paulo, R., Sacks, J., Cafeo, J. A., Cavendish, C.-H., and Tu, J. (2007a), 鈥淎 Framework for Validation of Computer Models,鈥? / Technometrics, 49, 138鈥?54. CrossRef
  4. Bayarri, M. J., Berger, J. O., Cafeo, J., Garcia-Donato, G., Liu, F., Palomo, J., Parthasarathy, R. J., Paulo, R., Sacks, J., and Walsh, D. (2007b), 鈥淐omputer Model Validation with Functional Output,鈥? / The Annals of Statistics, 35, 1874鈥?906. CrossRef
  5. Bliznyuk, N., Ruppert, D., and Shoemaker, C. A. (2011), 鈥淏ayesian Inference Using Efficient Interpolation of Computationally Expensive Densities With Variable Parameter Costs,鈥? / Journal of Computational and Graphical Statistics, 20, 636鈥?55. CrossRef
  6. 鈥?(2012), 鈥淟ocal Derivative-Free Approximation of Computationally Expensive Posterior Densities,鈥? / Journal of Computational and Graphical Statistics. doi:10.1080/10618600.2012.681255 .
  7. Bliznyuk, N., Ruppert, D., Shoemaker, C. A., Regis, R., Wild, S., and Mugunthan, P. (2008), 鈥淏ayesian Calibration and Uncertainty Analysis for Computationally Expensive Models Using Optimization and Radial Basis Function Approximation,鈥? / Journal of Computational and Graphical Statistics, 17, 270鈥?94. CrossRef
  8. Box, G. E. P., and Cox, D. R. (1964), 鈥淎n Analysis of Transformations,鈥? / Journal of the Royal Statistical Society, Series B, 26, 211鈥?46.
  9. Buhmann, M. D. (2003), / Radial Basis Functions, New York: Cambridge University Press. CrossRef
  10. Carroll, R. J., and Ruppert, D. (1984), 鈥淧ower Transformation When Fitting Theoretical Models to Data,鈥? / Journal of the American Statistical Association, 79, 321鈥?28. CrossRef
  11. 鈥?(1988), / Transformation and Weighting in Regression, New York: Chapman & Hall.
  12. Cumming, J. A., and Goldstein, M. (2009), 鈥淪mall Sample Bayesian Designs for Complex High-Dimensional Models Based on Information Gained Using Fast Approximations,鈥? / Journal of the American Statistical Association, 51, 377鈥?88.
  13. Eckhardt, K., Haverkamp, S., Fohrer, N., and Frede, H. G. (2002), 鈥淪WAT-G, A Version of SWAT99.2 Modified for Application to Low Mountain Range Catchments,鈥? / Physics and Chemistry of the Earth, 27, 641鈥?44. CrossRef
  14. Grizzetti, B., Bouraoui, F., Granlund, K., Rekolainen, S., and Bidoglio, G. (2003), 鈥淢odelling Diffuse Emission and Retention of Nutrients in the Vantaanjoki Watershed (Finland) Using the SWAT Model,鈥? / Ecological Modelling, 169, 25鈥?8. CrossRef
  15. Hamilton, J. D. (1994), / Time Series Analysis, Princeton: Princeton University Press.
  16. Higdon, D., Kennedy, M., Cavendish, J. C., Cafeo, J. A., and Ryne, R. D. (2004), 鈥淐ombining Field Data and Computer Simulations for Calibration and Prediction,鈥? / SIAM Journal of Scientific Computation, 26, 448鈥?66. CrossRef
  17. Kennedy, M. C., and O鈥橦agan, A. (2001), 鈥淏ayesian Calibration of Computer Models,鈥? / Journal of the Royal Statistical Society, Series B, 63, 425鈥?64. CrossRef
  18. Levy, S., and Steinberg, D. M. (2010), 鈥淐omputer Experiments: A Review,鈥? / AStA Advances in Statistical Analysis, 94, 311鈥?24. CrossRef
  19. Qian, P. Z. G. (2009), 鈥淣ested Latin Hypercube Design,鈥? / Biometrika, 96, 957鈥?70. CrossRef
  20. Qian, P. Z. G., and Wu, C. F. J. (2008), 鈥淏ayesian Hierarchical Modeling for Integrating Low-Accuracy and High-Accuracy Experiments,鈥? / Technometrics, 50, 192鈥?04. CrossRef
  21. Rasmussen, C. E. (2003), ````Gaussian Processes to Speed up Hybrid Monte Carlo for Expensive Bayesian Integrals,鈥?in / Bayesian Statistics 7, eds. J. M. Bernardo, J. O. Berger, A. P. Berger, and A. F. M. Smith, pp. 651鈥?59.
  22. Regis, R. G., and Shoemaker, C. A. (2009), 鈥淧arallel Stochastic Global Optimization Using Radial Basis Functions,鈥? / INFORMS Journal on Computing, 21, 411鈥?26. CrossRef
  23. Ruppert, D., Wand, M.聽P., and Carroll, R.聽J. (2003), / Semiparametric Regression, Cambridge: Cambridge University Press. CrossRef
  24. Santner, T. J., Williams, B. J., and Notz, W. I. (2010), / The Design and Analysis of Computer Experiments, New York: Springer.
  25. Shoemaker, C. A., Regis, R., and Fleming, R. (2007), 鈥淲atershed Calibration Using Multistart Local Optimization and Evolutionary Optimization With Radial Basis Function Approximation,鈥? / Journal of Hydrologic Science, 52, 450鈥?65. CrossRef
  26. Singh, A. (2011), 鈥淕lobal Optimization of Computationally Expensive Hydrologic Simulation Models,鈥?Ph.D. Thesis in Civil and Environmental Engineering, Cornell University.
  27. Tierney, L. (1994), 鈥淢arkov Chains for Exploring Posterior Distributions,鈥? / The Annals of Statistics, 22, 1701鈥?786. CrossRef
  28. Tierney, L., and Kadane, J. (1986), 鈥淎ccurate Approximations for Posterior Moments and Marginal Densities,鈥? / Journal of the American Statistical Association, 81, 82鈥?6. CrossRef
  29. Tjelmeland, H., and Hegstad, B. K. (2001), 鈥淢odel Jumping Proposals in MCMC,鈥? / Scandinavian Journal of Statistics, 28, 205鈥?23. CrossRef
  30. Tolson, B. (2005), 鈥淎utomatic Calibration, Management and Uncertainty Analysis: Phosphorus Transport in the Cannonsville Watershed, 鈥?Ph.D. dissertation, School of Civil and Environmental Engineering, Cornell University.
  31. Tolson, B., and Shoemaker, C. A. (2004), 鈥淲atershed Modeling of the Cannonsville Basin Using SWAT2000: Model Development, Calibration and Validation for the Prediction of Flow, Sediment and Phosphorus Transport to the Cannonsville Reservoir, Version 1,鈥?Technical Report, School of Civil and Environmental Engineering, Cornell University. Available at http://ecommons.library.cornell.edu/handle/1813/2710.
  32. 鈥?(2007a), 鈥淭he Dynamically Dimensioned Search Algorithm for Computationally Efficient Automatic Calibration of Environmental Simulation Models,鈥? / Water Resources Research, 43, W01413. doi:10.1029/2005WR004723 . CrossRef
  33. 鈥?(2007b), 鈥淐annonsville Reservoir Watershed SWAT2000 Model Development, Calibration and Validation,鈥? / Journal of Hydrology, 337, 68鈥?9 doi:10.1016/j.jhydrol.2007.01.017 . CrossRef
  34. Vanden Berghen, F., and Bersini, H. (2005), 鈥淐ONDOR, a New Parallel, Constrained Extension of Powell鈥檚 UOBYQA Algorithm: Experimental Results and Comparison With the DFO Algorithm,鈥? / Journal of Computational and Applied Mathematics, 181, 157鈥?75. CrossRef
  35. Wild, S. M., and Shoemaker, C. A. (2011), 鈥淕lobal Convergence of Radial Basis Functions Trust Region Derivative-Free Algorithms,鈥? / SIAM Journal on Optimization, 20, 387鈥?15.
  36. Wild, S. M., Regis, R. G., and Shoemaker, C. A. (2007), 鈥淥RBIT: Optimization by Radial Basis Function Interpolation in Trust-Regions,鈥? / SIAM Journal on Scientific Computing, 30, 3197鈥?219. CrossRef
  
• 作者单位：David Ruppert (1)
  Christine A. Shoemaker (2)
  Yilun Wang (3) (4)
  Yingxing Li (5)
  Nikolay Bliznyuk (6)
  
  1. School of Operations Research and Information Engineering and Department of Statistical Science, Cornell University, Comstock Hall, Ithaca, NY, 14853, USA
  2. School of Civil and Environmental Engineering and School of Operations Research and Information Engineering, Cornell University, Hollister Hall, Ithaca, NY, 14853, USA
  3. School of Mathematical Sciences, University of Electronic Science and Technology of China, Chengdu, Sichuan, 611731, China
  4. School of Civil and Environmental Engineering, Cornell University, Ithaca, NY, 14853, USA
  5. Wang Yanan Institute for Studies in Economics, Xiamen University, Xiamen, China
  6. Department of Statistics (IFAS), University of Florida, Gainesville, FL, 32611, USA
  
• ISSN：1537-2693

文摘

Bayesian MCMC calibration and uncertainty analysis for computationally expensive models is implemented using the SOARS (Statistical and Optimization Analysis using Response Surfaces) methodology. SOARS uses a radial basis function interpolator as a surrogate, also known as an emulator or meta-model, for the logarithm of the posterior density. To prevent wasteful evaluations of the expensive model, the emulator is built only on a high posterior density region (HPDR), which is located by a global optimization algorithm. The set of points in the HPDR where the expensive model is evaluated is determined sequentially by the GRIMA algorithm described in detail in another paper but outlined here. Enhancements of the GRIMA algorithm were introduced to improve efficiency. A case study uses an eight-parameter SWAT2005 (Soil and Water Assessment Tool) model where daily stream flows and phosphorus concentrations are modeled for the Town Brook watershed which is part of the New York City water supply. A Supplemental Material file available online contains additional technical details and additional analysis of the Town Brook application.