Refined Stratified Sampling for efficient Monte Carlo based uncertainty quantification

Abstract A general adaptive approach rooted in stratified sampling (SS) is proposed for sample-based uncertainty quantification (UQ). To motivate its use in this context the space-filling, orthogonality, and projective properties of SS are compared with simple random sampling and Latin hypercube sampling (LHS). SS is demonstrated to provide attractive properties for certain classes of problems. The proposed approach, Refined Stratified Sampling (RSS), capitalizes on these properties through an adaptive process that adds samples sequentially by dividing the existing subspaces of a stratified design. RSS is proven to reduce variance compared to traditional stratified sample extension methods while providing comparable or enhanced variance reduction when compared to sample size extension methods for LHS – which do not afford the same degree of flexibility to facilitate a truly adaptive UQ process. An initial investigation of optimal stratification is presented and motivates the potential for major advances in variance reduction through optimally designed RSS. Potential paths for extension of the method to high dimension are discussed. Two examples are provided. The first involves UQ for a low dimensional function where convergence is evaluated analytically. The second presents a study to asses the response variability of a floating structure to an underwater shock.

[1]  P. Hellekalek,et al.  Random and Quasi-Random Point Sets , 1998 .

[2]  Paul Glasserman,et al.  Monte Carlo Methods in Financial Engineering , 2003 .

[3]  Michael A. Sprague,et al.  Spectral elements and field separation for an acoustic fluid subject to cavitation , 2003 .

[4]  Richard J. Beckman,et al.  A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output From a Computer Code , 2000, Technometrics.

[5]  David Hinkley,et al.  Bootstrap Methods: Another Look at the Jackknife , 2008 .

[6]  A. Owen.  A Central Limit Theorem for Latin Hypercube Sampling , 1992 .

[7]  William H. Press,et al.  Recursive stratified sampling for multidimensional Monte Carlo integration , 1990 .

[8]  Thomas W. Lucas,et al.  Efficient Nearly Orthogonal and Space-Filling Latin Hypercubes , 2007, Technometrics.

[9]  Michael A. Sprague,et al.  Computational Treatments of Cavitation Effects in Near-Free-Surface Underwater Shock Analysis , 2001 .

[10]  Peter Winker,et al.  Centered L2-discrepancy of random sampling and Latin hypercube design, and construction of uniform designs , 2002, Math. Comput..

[11]  Michael D. Shields,et al.  Targeted random sampling: a new approach for efficient reliability estimation for complex systems , 2015 .

[12]  M. E. Johnson,et al.  Minimax and maximin distance designs , 1990 .

[13]  Mircea Grigoriu,et al.  Reduced order models for random functions. Application to stochastic problems , 2009 .

[14]  R. Iman,et al.  A distribution-free approach to inducing rank correlation among input variables , 1982 .

[15]  F. J. Hickernell.  Lattice rules: how well do they measure up? in random and quasi-random point sets , 1998 .

[16]  J. Hammersley.  SIMULATION AND THE MONTE CARLO METHOD , 1982 .

[17]  Miroslav Vorechovský,et al.  Hierarchical Refinement of Latin Hypercube Samples , 2015, Comput. Aided Civ. Infrastructure Eng..

[18]  Peter Z. G. Qian.  Nested Latin hypercube designs , 2009 .

[19]  A. Olsson,et al.  On Latin hypercube sampling for structural reliability analysis , 2003 .

[20]  I. S. Sandler,et al.  Uncoupling Approximations in Fluid-Structure Interaction Problems With Cavitation , 1981 .

[21]  Saltelli Andrea,et al.  Global Sensitivity Analysis: The Primer , 2008 .

[22]  G. Lepage.  A new algorithm for adaptive multidimensional integration , 1978 .

[23]  V. R. Joseph,et al.  ORTHOGONAL-MAXIMIN LATIN HYPERCUBE DESIGNS , 2008 .

[24]  T. J. Mitchell,et al.  Exploratory designs for computational experiments , 1995 .

[25]  Michael J. Gilman,et al.  A brief survey of stopping rules in Monte Carlo simulations , 1968 .

[26]  Franz Aurenhammer,et al.  Voronoi diagrams—a survey of a fundamental geometric data structure , 1991, CSUR.

[27]  Christopher J. Roy,et al.  A comprehensive framework for verification, validation, and uncertainty quantification in scientific computing , 2011 .

[28]  K. D. Tocher,et al.  The art of simulation , 1967 .

[29]  Runze Li,et al.  Design and Modeling for Computer Experiments , 2005 .

[30]  M. Stein.  Large sample properties of simulations using latin hypercube sampling , 1987 .

[31]  Jeong‐Soo Park.  Optimal Latin-hypercube designs for computer experiments , 1994 .

[32]  M. Liefvendahl,et al.  A study on algorithms for optimization of Latin hypercubes , 2006 .

[33]  J. Hammersley,et al.  Monte Carlo Methods , 1965 .

[34]  Hans H. Bleich,et al.  Interaction between structures and bilinear fluids , 1970 .

[35]  Chang-Xing Ma,et al.  Wrap-Around L2-Discrepancy of Random Sampling, Latin Hypercube and Uniform Designs , 2001, J. Complex..

[36]  FangKai-Tai,et al.  Centered L2-discrepancy of random sampling and Latin hypercube design, and construction of uniform designs , 2002 .

[37]  Kenny Q. Ye.  Orthogonal Column Latin Hypercubes and Their Application in Computer Experiments , 1998 .

[38]  Adam Hapij,et al.  Mapping model validation metrics to subject matter expert scores for model adequacy assessment , 2014, Reliab. Eng. Syst. Saf..

[39]  Drahomír Novák,et al.  Sample size extension in stratified sampling: Theory and software implementation , 2014 .

[40]  Aleš Florian,et al.  An efficient sampling scheme: Updated Latin Hypercube Sampling , 1992 .

[41]  George N. Karystinos,et al.  Fast generation of space-filling Latin hypercube sample designs. , 2010 .

[42]  Michael D. McKay,et al.  Evaluating Prediction Uncertainty , 1995 .

[43]  Boxin Tang.  Orthogonal Array-Based Latin Hypercubes , 1993 .

[44]  D. Huntington,et al.  Improvements to and limitations of Latin hypercube sampling , 1998 .

[45]  Kenny Q. Ye,et al.  Algorithmic construction of optimal symmetric Latin hypercube designs , 2000 .

[46]  C. F. Jeff Wu,et al.  Experiments: Planning, Analysis, and Parameter Design Optimization , 2000 .

[47]  G. G. Wang,et al.  Adaptive Response Surface Method Using Inherited Latin Hypercube Design Points , 2003 .

[48]  M. Kalos,et al.  Monte Carlo methods , 1986 .

[49]  Charles Tong,et al.  Refinement strategies for stratified sampling methods , 2006, Reliab. Eng. Syst. Saf..

[50]  Wei Chen,et al.  Optimizing Latin hypercube design for sequential sampling of computer experiments , 2009 .

[51]  Jon C. Helton,et al.  Extension of Latin hypercube samples with correlated variables , 2008, Reliab. Eng. Syst. Saf..

[52]  D. Novák,et al.  CORRELATION CONTROL IN SMALL-SAMPLE MONTE CARLO TYPE SIMULATIONS I: A SIMULATED ANNEALING APPROACH , 2009 .

[53]  G. Rennen,et al.  Nested maximin Latin hypercube designs , 2009 .

[54]  Hans Janssen,et al.  Monte-Carlo based uncertainty analysis: Sampling efficiency and sampling convergence , 2013, Reliab. Eng. Syst. Saf..