Asymptotic Analysis of Multilevel Best Linear Unbiased Estimators

We study the computational complexity and variance of multilevel best linear unbiased estimators introduced in [D. Schaden and E. Ullmann, SIAM/ASA J. Uncert. Quantif., (2020)]. We specialize the results in this work to PDE-based models that are parameterized by a discretization quantity, e.g., the finite element mesh size. In particular, we investigate the asymptotic complexity of the so-called sample allocation optimal best linear unbiased estimators (SAOBs). These estimators have the smallest variance given a fixed computational budget. However, SAOBs are defined implicitly by solving an optimization problem and are difficult to analyze. Alternatively, we study a class of auxiliary estimators based on the Richardson extrapolation of the parametric model family. This allows us to provide an upper bound for the complexity of the SAOBs, showing that their complexity is optimal within a certain class of linear unbiased estimators. Moreover, the complexity of the SAOBs is not larger than the complexity of Multilevel Monte Carlo. The theoretical results are illustrated by numerical experiments with an elliptic PDE.

[1]  Benjamin Peherstorfer,et al.  Survey of multifidelity methods in uncertainty propagation, inference, and optimization , 2018, SIAM Rev..

[2]  Gilles Pagès,et al.  Multilevel Richardson-Romberg Extrapolation , 2014 .

[3]  F. Chatelin.  Spectral approximation of linear operators , 2011 .

[4]  Stephen P. Boyd,et al.  Convex Optimization , 2004, Algorithms and Theory of Computation Handbook.

[5]  Benjamin Peherstorfer,et al.  Convergence analysis of multifidelity Monte Carlo estimation , 2018, Numerische Mathematik.

[6]  Stephen S. Lavenberg,et al.  A Perspective on the Use of Control Variables to Increase the Efficiency of Monte Carlo Simulations , 1981 .

[7]  K. A. Cliffe,et al.  Multilevel Monte Carlo methods and applications to elliptic PDEs with random coefficients , 2011, Comput. Vis. Sci..

[8]  R. Rannacher,et al.  Asymptotic error expansion and Richardson extranpolation for linear finite elements , 1986 .

[9]  Frédéric Vrins,et al.  An Antithetic Approach of Multilevel Richardson-Romberg Extrapolation Estimator for Multidimensional SDES , 2016, NAA.

[10]  J. Stoer,et al.  Numerical treatment of ordinary differential equations by extrapolation methods , 1966 .

[11]  L. Richardson.  The Approximate Arithmetical Solution by Finite Differences of Physical Problems Involving Differential Equations, with an Application to the Stresses in a Masonry Dam , 1911 .

[12]  Claude Brezinski,et al.  Convergence acceleration during the 20th century , 2000 .

[13]  D. Talay,et al.  Expansion of the global error for numerical schemes solving stochastic differential equations , 1990 .

[14]  Gianluca Geraci,et al.  A Generalized Framework for Approximate Control Variates , 2018 .

[15]  Wolfgang L. Wendland,et al.  A new approach to Richardson extrapolation in the finite element method for second order elliptic problems , 2009, Math. Comput..

[16]  Michael B. Giles,et al.  Multilevel Monte Carlo Path Simulation , 2008, Oper. Res..

[17]  Michael L. Stein,et al.  Interpolation of spatial data , 1999 .

[18]  Benjamin Peherstorfer,et al.  Optimal Model Management for Multifidelity Monte Carlo Estimation , 2016, SIAM J. Sci. Comput..

[19]  Michael B. Giles,et al.  Multilevel Monte Carlo methods , 2013, Acta Numerica.

[20]  Gianluca Geraci,et al.  A generalized approximate control variate framework for multifidelity uncertainty quantification , 2018, J. Comput. Phys..

[21]  Elisabeth Ullmann,et al.  On Multilevel Best Linear Unbiased Estimators , 2020, SIAM/ASA J. Uncertain. Quantification.

[22]  Tony Shardlow,et al.  Improving multilevel Monte Carlo for stochastic differential equations with application to the Langevin equation , 2015, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences.