Goal-oriented adaptive surrogate construction for stochastic inversion

Abstract For computationally expensive models, surrogate response surfaces are often employed to increase the number of samples used in approximating the solution to a stochastic inverse problem. The result is generally a trade-off in errors where the stochastic error is reduced at the cost of an increase in deterministic/discretization errors in the evaluation of the surrogate. Such stochastic errors pollute predictions based on the stochastic inverse. In this work, we formulate a method for adaptively creating a special class of surrogate response surfaces with these sources of error in mind. Adjoint techniques are used to enhance the local approximation properties of the surrogate allowing the construction of a higher-level enhanced surrogate. Using these two levels of surrogates, appropriately derived local error indicators are computed and used to guide refinement of both levels of the surrogates. Three types of refinement strategies are presented and combined in an iterative adaptive surrogate construction algorithm. Numerical examples, including a complex vibroacoustics application, demonstrate how this adaptive strategy allows for accurate predictions under uncertainty for a much smaller computational cost than uniform refinement.

[1]  Habib N. Najm,et al.  Stochastic spectral methods for efficient Bayesian solution of inverse problems , 2005, J. Comput. Phys..

[2]  D. Estep.  A posteriori error bounds and global error control for approximation of ordinary differential equations , 1995 .

[3]  Andrew M. Stuart,et al.  Inverse problems: A Bayesian perspective , 2010, Acta Numerica.

[4]  J. Oden,et al.  A Posteriori Error Estimation in Finite Element Analysis , 2000 .

[5]  Barbara Wohlmuth,et al.  Reduced Basis Isogeometric Mortar Approximations for Eigenvalue Problems in Vibroacoustics , 2016, 1606.01163.

[6]  O. L. Maître,et al.  Spectral Methods for Uncertainty Quantification: With Applications to Computational Fluid Dynamics , 2010 .

[7]  Tim Wildey,et al.  A Posteriori Error Analysis of Parameterized Linear Systems Using Spectral Methods , 2012, SIAM J. Matrix Anal. Appl..

[8]  Jinglai Li,et al.  Adaptive Construction of Surrogates for the Bayesian Solution of Inverse Problems , 2013, SIAM J. Sci. Comput..

[9]  Rolf Rannacher,et al.  An optimal control approach to a posteriori error estimation in finite element methods , 2001, Acta Numerica.

[10]  Tim Wildey,et al.  Error Decomposition and Adaptivity for Response Surface Approximations from PDEs with Parametric Uncertainty , 2015, SIAM/ASA J. Uncertain. Quantification.

[11]  Charles J. Geyer,et al.  Practical Markov Chain Monte Carlo , 1992 .

[12]  M. Giles,et al.  Adjoint methods for PDEs: a posteriori error analysis and postprocessing by duality , 2002, Acta Numerica.

[13]  T. Başar,et al.  A New Approach to Linear Filtering and Prediction Problems , 2001 .

[14]  I. Babuska,et al.  A‐posteriori error estimates for the finite element method , 1978 .

[15]  Tim Wildey,et al.  Propagation of Uncertainties Using Improved Surrogate Models , 2012, SIAM/ASA J. Uncertain. Quantification.

[16]  J. Tinsley Oden,et al.  Solution verification, goal-oriented adaptive methods for stochastic advection-diffusion problems , 2010 .

[17]  Nicholas Zabaras,et al.  An efficient Bayesian inference approach to inverse problems based on an adaptive sparse grid collocation method , 2009 .

[18]  Anders Logg,et al.  The FEniCS Project Version 1.5 , 2015 .

[19]  Patrick R. Conrad,et al.  Accelerating Asymptotically Exact MCMC for Computationally Intensive Models via Local Approximations , 2014, 1402.1694.

[20]  Mohamed S. Ebeida,et al.  VPS: VORONOI PIECEWISE SURROGATE MODELS FOR HIGH-DIMENSIONAL DATA FITTING , 2017 .

[21]  Anders Logg,et al.  Automated Solution of Differential Equations by the Finite Element Method: The FEniCS Book , 2012 .

[22]  Fabio Nobile,et al.  A Sparse Grid Stochastic Collocation Method for Partial Differential Equations with Random Input Data , 2008, SIAM J. Numer. Anal..

[23]  Hoon Kim,et al.  Monte Carlo Statistical Methods , 2000, Technometrics.

[24]  R. Ghanem,et al.  Uncertainty propagation using Wiener-Haar expansions , 2004 .

[25]  Craig B. Borkowf,et al.  Random Number Generation and Monte Carlo Methods , 2000, Technometrics.

[26]  Serge Prudhomme,et al.  Adaptive surrogate modeling for response surface approximations with application to bayesian inference , 2015, Adv. Model. Simul. Eng. Sci..

[27]  Peter Green,et al.  Markov chain Monte Carlo in Practice , 1996 .

[28]  Alexander Düster,et al.  Book Review: Leszek Demkowicz, Computing with hp‐adaptive finite elements, Volume 1, One and two dimensional elliptic and Maxwell problems , 2007 .

[29]  Endre Süli,et al.  Adaptive finite element methods for differential equations , 2003, Lectures in mathematics.

[30]  Dongbin Xiu,et al.  The Wiener-Askey Polynomial Chaos for Stochastic Differential Equations , 2002, SIAM J. Sci. Comput..

[31]  Geir Evensen,et al.  The Ensemble Kalman Filter: theoretical formulation and practical implementation , 2003 .

[32]  Qiqi Wang,et al.  Erratum: Active Subspace Methods in Theory and Practice: Applications to Kriging Surfaces , 2013, SIAM J. Sci. Comput..

[33]  Omri Rand,et al.  Analytical Methods in Anisotropic Elasticity: with Symbolic Computational Tools , 2004 .

[34]  R. Ghanem,et al.  Multi-resolution analysis of wiener-type uncertainty propagation schemes , 2004 .

[35]  David M. Gay,et al.  Automatic Differentiation of C++ Codes for Large-Scale Scientific Computing , 2006, International Conference on Computational Science.

[36]  Andrew M. Stuart,et al.  Approximation of Bayesian Inverse Problems for PDEs , 2009, SIAM J. Numer. Anal..

[37]  George E. Karniadakis,et al.  Beyond Wiener–Askey Expansions: Handling Arbitrary PDFs , 2006, J. Sci. Comput..

[38]  Troy Butler,et al.  A Measure-Theoretic Interpretation of Sample Based Numerical Integration with Applications to Inverse and Prediction Problems under Uncertainty , 2017, SIAM J. Sci. Comput..

[39]  Tsuyoshi Murata,et al.  {m , 1934, ACML.

[40]  Donald J. Estep,et al.  Fast and reliable methods for determining the evolution of uncertain parameters in differential equations , 2006, J. Comput. Phys..