betaFeedback

WorksMathematics

Triple Products of Correlated Multivariate Hermite Polynomials

Abstract

Stochastic Galerkin schemes (SGS) are used to approximate solutions to differential equations with stochastic parameters. In particular, they express the solution with an infinite orthogonal functional basis of the parameter space and derive a deterministic surrogate system of PDEs whose solution approximates that of the original stochastic problem. When the parameters are correlated, the infinite series representation of the solution does not provably converge for all joint distributions; further, classic decorrelating transformations (e.g., Rosenblatt, Nataf) often introduce known numerical instabilities. Rahman (2017) established convergence in probability and distribution for correlated Gaussian parameters when the basis functionals are multivariate Hermite polynomials (MHPs), where these polynomials are only weakly orthogonal with respect to the underlying joint density. Due to weak orthogonality, the NNth moments of MHPs that emerge during Galerkin projection do not trivially decompose into products of one-dimensional expectations via tensorization. Namely, an NNth-order moment has no closed-form expression for N>2N > 2 and consequently is estimated numerically. Because the number of such moments grows combinatorially with parameter dimension and polynomial degree, approximating each one individually via sampling or quadrature can quickly become cost-prohibitive. In this paper, we introduce and prove several foundational properties of MHPs. These results culminate in an explicit formula for evaluating the N=3N = 3 moments (triple products) directly without numerical estimation.

Introduction

This paper considers a subset of forward propagation problems, in which the goal is to quantify the variability in some output quantity of interest (QoI) given uncertain model parameters represented by random variables. The random variables are prescribed a joint probability density function, which is selected by the scientist or engineer based on physical knowledge of the problem. Using the notation in Figure 1, we can imagine that the model functional F\mathcal{F} represents a system of partial differential equations (PDEs) with solution uu, where the uncertain input parameters appear in initial or boundary conditions. Accordingly, the response uu is a stochastic PDE solution, since it is a function of the random parameters.

Framework for uncertainty propagation problems

Figure 1. Framework for uncertainty propagation problems.

Several methods have been developed within the field of uncertainty quantification1 (UQ) to tackle the forward propagation of uncertainty, especially in the context of solving stochastic systems in engineering applications [5, 6, 30, 36, 44, 45, 53]. As illustrated in Figure 2, these methods in UQ can be classified into three broad categories: sampling methods (subsection 1.1), quadrature and stochastic collocation, and spectral Galerkin methods (subsection 1.2). While this paper does not discuss quadrature and stochastic collocation, there is a wealth of research and literature on the subject [22, 28, 56, 64].

The three pillars of UQ for analyzing forward propagation problems

Figure 2. The three pillars of UQ for analyzing forward propagation problems.

Sampling methods

The sampling approach uses many realizations of an input random variable to determine distributional information about the output solution of interest. The most common example is a simple Monte Carlo (MC) simulation. The advantages of simple MC are (1) it is generally straight-forward to implement, and (2) its accuracy is not dependent on the dimensionality of the uncertain input parameter (but instead on the number of parameter realizations) [15, 41]. However, by the law of large numbers, MC algorithms suffer from a slow convergence rate proportional to the inverse square root of the number of samples [32, 44, 56]. Accordingly, if a single sample evaluation of the output solution of interest is time consuming, and/or a large number of sample evaluations is required to converge within a desired accuracy level, MC simulations can become prohibitively expensive. Further, if the goal is to assess the behavior of the stochastic solution given low-probability events, sampling methods are not ideal, since they will require a large number of realizations [63].

To accelerate convergence, several techniques have been developed, such as Latin Hypercube Sampling, Quasi-Monte Carlo (QMC), and Markov Chain Monte Carlo (MCMC) [8, 29, 56]; for an excellent overview, see [41, 42]. These methods can improve the efficiency of the brute-force Monte Carlo method — though additional restrictions are imposed based on their specific designs, meaning their applicability may be limited in some contexts [63].

Stochastic Galerkin schemes (SGS)

Stochastic Galerkin schemes (SGS). For MC algorithms, if each sample evaluation is expensive — as is often the case for a PDE solution underlying a large-scale simulation — the slow convergence can make obtaining tens of thousands of evaluations computationally infeasible [5, 14, 38]. These computational constraints are worsened when the resolved model requires a high degree of accuracy. In such cases, stochastic Galerkin methods might be better suited to the problem at hand.

Stochastic Galerkin schemes (SGS) are used to approximate the solution to parametrized differential equations. In particular, they utilize an infinite functional basis of the parameter space to express the solution and derive a deterministic system of PDEs, which can then be solved with standard numerical techniques [4, 17, 56]. The deterministic coefficients in the functional series representation of the solution are sometimes called random modes. Since Galerkin methods are based on projecting the residual onto the space of approximating polynomials, its accuracy is optimal in an L2L^2 sense [56]. SGS are part of a broader class known as spectral methods [1, 5, 17, 18, 34, 44].

SGS are called intrusive in the sense that a new system associated with the projection is developed and implemented, where the new system requires constructing expectations or inner products not typically used in deterministic codes; therefore, existing codes cannot be used directly to recover the random modes [6, 44, 56]. In other words, intrusive methods for uncertainty quantification reformulate the original stochastic model equations to obtain a system of deterministic equations — the trade-off being that the deterministic system involves more equations and variables than the original model. For applications with large legacy codes or ‘black-box’ executable files, a non-intrusive approach (e.g., Monte Carlo, stochastic collocation/quadrature [64], etc.) might be more appropriate [5].

In the conventional approach for SGS, the functional basis elements are chosen to be polynomials that are orthogonal with respect to the joint density of the random input variables. This representation has its roots in the work of Wiener [62], who expressed a Gaussian process as a countably infinite series of Hermite polynomials. Ghanem and Spanos [18] truncated Wiener’s representation and used the resulting finite series as a key ingredient in a stochastic finite element method. SGS based on polynomial expansions are often referred to as polynomial chaos approaches.

When compared to MC algorithms, initial spectral method applications to UQ problems showed orders-of-magnitude reductions in the cost needed to estimate statistics with comparable accuracy [66]; however, there are of course some exceptions, e.g., see [3, 16]. Several PCE extensions have been developed in the last years [38] that have shown to be effective in a large number of applications in different fields: fluid dynamics [14, 34, 40, 44], air-water flows in soil [57], aircraft design [47], radiation therapy [43, 51], etc. Further, spectral Galerkin methods have the advantage of returning the entire distribution of the output QoI\mathrm{QoI}, from which any descriptive statistics (mean, variance, etc.) can be computed directly; this is in direct contrast to sampling strategies, which only compute pre-specified descriptive statistics.

Limitations of polynomial chaos

Limitations of polynomial chaos. Classically, polynomial chaos methods have the following disadvantages.

Distribution of stochastic input parameter. In this paper, we assume that the stochastic random variable input ξ\xi is Gaussian. However, in a series of papers [26, 63, 65, 66], Xiu et al. introduced spectral expansions for non-Gaussian ξ\xi, where the corresponding basis polynomials are selected to be orthogonal with respect to the density of ξ\xi. For example, Legendre polynomials are selected when ξ∼U(a,b)\xi\sim U(a,b), Laguerre polynomials when ξ∼Γ(a,b)\xi\sim\Gamma(a,b), etc. This spearheaded the use of the so-called generalized polynomial chaos expansion (gPCE) in an explosion of research applications [12, 27, 45].

With regard to the convergence of these generalized expansions, the issue was approached rather guardedly [13]; one of the earlier papers [65] on the topic states, “Convergence to second-order stochastic processes can be possibly obtained as a generalization of Cameron–Martin theorem.” Only in 2010 did Ernst et al. prove that an arbitrary random variable with finite variance can be expressed as a generalized polynomial chaos expansion only if the underlying probability measure is uniquely determined by its moments [13] — a condition which fails in some simple cases, such as when ξ\xi has a lognormal distribution (see pages 106 - 107 of [11]). Restrictions like this preclude the use of a polynomial chaos expansion for general densities constructed using Bayesian techniques [56].

Poor convergence for non-smooth solutions. As noted by Xiu et al., the exponential convergence rate of the PCE degrades when mapping the random input vector to the response is not sufficiently smooth [63, 66] due to Gibbs phenomenon [20, 23]. Further, Gibbs oscillations may cause a spectral approximation to seemingly attain nonphysical values. In the worst case, the PCE may fail to capture essential features of the stochastic quantity of interest [35, 44]. Consequently, the PDEs considered in this paper are restricted to those whose solutions are smooth within the domain of interest.

Assumed independence of stochastic parameters. For systems that depend on multiple stochastic parameters, we can extend the univariate versions of the Wiener-Hermite PCE, the gPCE, and stochastic Galerkin schemes (SGS) in general. In the multivariate case, one of the essential assumptions when applying SGS is that the stochastic model parameters are independent. This assumption allows the joint distribution of the random variables to decompose as a product of probability densities, enabling the easy construction of multivariate orthogonal polynomials via tensorizing the spaces of univariate orthogonal polynomials [46, 56, 63]. These orthogonal basis functionals are consistent with the underlying measure [13] and thus constitute a complete Hilbert basis [46]; as clarified only recently [13], this completeness2 is vitally important for convergence of the truncated PCE to the correct limit.

Curse of dimensionality. As with several numerical techniques, polynomial chaos suffers from the curse of dimensionality, which is to say that as the number of parameters increases, the work required to compute the approximation increases combinatorially [6, 14, 46, 56, 63]. Over the years, there has been a vast amount of literature3 on methods for alleviating this curse of dimensionality for certain classes of problems, several of which involve differential equations with stochastic inputs [5]. Even so, PC methods are typically used to address low-dimensional problems — that is, problems for which the number of parameters is not very large (say, less than ten) [46].

Multivariate polynomial chaos

1.4. Multivariate polynomial chaos. As described in subsection 1.3, in the multivariate case, the PCE (along with the gPCE and SGS overall) is founded on the assumption that its input random variables are independent [5, 14, 38, 44, 56].

Decorrelating transformations. In reality, there may exist significant correlation or dependence among input variables.6 One approach is to apply a decorrelating transformation [14, 38, 56] and work with the PCE in the resulting independent7 parameters; specifically, some options involve linear decorrelating maps (e.g., Rosenblatt [48], Nataf [37]), the K-L expansion [28], or proper orthogonal decomposition [31].

The challenge is that such transformations tend to introduce discontinuities or steep gradients [14, 44, 66], which (as discussed in subsection 1.3) may cause a significant slowdown in the convergence speed of PC methods due to Gibbs phenomenon [20, 23, 45]. Further, for applications like general nonlinear conservation laws (e.g., fluid dynamics problems) [45], diffusion modeling in multilayered materials [14], and in blood flow simulations at arterial junctions [24], the PDE solution may have discontinuities and/or steep gradients to begin with — so these effects on convergence are worsened by decorrelating; the result is an expansion with a lower convergence rate as the polynomial degree increases [12, 14, 20, 23, 63, 66]. Still, utilizing such transformations is an active area of research in the context of polynomial chaos [14, 25, 61].

The work in this paper does not use transformation-based methods.

Transformation-free approaches. There have been several transformation-free approaches to account for arbitrary but dependent Gaussian random variables. Using Gram–Schmidt orthogonalization [19], Navarro et al. discussed construction of a PCE with multivariate orthogonal polynomials for correlated variables [38]. However, the existence of multivariate Hermite orthogonal polynomials, which can be used as a basis for dependent Gaussian measures, has not been recognized [46]. Soize and Ghanem proposed orthogonal functional bases with regard to a general dependent probability measure of random input, but the bases elements are not necessarily polynomials [58]; as a result, an analytical treatment of PCE-based statistics in this case is highly non-trivial [56]. Furthermore, both the works of Navarro and Soize/Ghanem do not address a fundamental concern raised by Ernst et al. about the completeness of their orthogonal basis functionals, which was shown to be a necessary condition for convergence [13, 46].

In 2017, Rahman constructed a PCE with weakly orthogonal (defined in (5)) basis functionals consistent with the Gaussian measure, which was shown to converge to the correct limit in mean-square, probability, and distribution [46]. We will refer to this series expansion as the correlated multivariate PCE (or CPCE)8 and use it exclusively throughout the rest of this document.

Higher-order products

In the Galerkin framework, one projects weighted residuals onto a finite-dimensional subspace spanned by basis functionals to provide the constraints required to solve for the deterministic coefficients [56]. When utilizing the CPCE, the projection step introduces expectations of products of multivariate Hermite polynomials in the Gaussian parameters. One major consequence of weak orthogonality is that these NNth products of the form Eξ(Hα(1)(ξ)⋯Hα(N)(ξ))\mathbb{E}_{\boldsymbol{\xi}}\left(H_{\boldsymbol{\alpha}^{(1)}}(\boldsymbol{\xi}) \cdots H_{\boldsymbol{\alpha}^{(N)}}(\boldsymbol{\xi})\right) no longer decompose into a multiplication of their one-dimensional counterparts. Often these expectation integrals over the nn-dimensional parameter space are approximated numerically — and doing so can be highly nontrivial [5, 14, 56]. Typically, these integrals are estimated with some combination of sampling methods, standard Gaussian quadrature/collocation, and/or sparse grid techniques (when possible) [1, 22, 46, 54, 63, 64] — that is, with the statistical methods discussed in subsection 1.1. In 2017, Rahman first provided a formula for the double products (N=2N = 2) or second-order moments, Eξ(Hα(ξ)Hβ(ξ))\mathbb{E}_{\xi}(H_{\alpha}(\xi)H_{\beta}(\xi)), that appear during Galerkin projection [46]. Prior to the work given in this paper, in the presence of multiple correlated uncertainties, there was no closed-form expression for the NNth products of the multivariate Hermite polynomials for any NN larger than two.

Summary of contributions

This paper presents technical contributions regarding multivariate Hermite polynomials (MHPs) whose inputs are correlated Gaussian random variables, culminating in a formula for their triple products. We assume ξ∼N(0,Σ)\xi\sim\mathcal{N}(0,\Sigma) for known (SPD) covariance matrix (Σ\Sigma). All expected values whose inputs involve MHPs are understood to mean expectation with respect to the dξd\xi measure. The key results are outlined below.

  • An MHP indexed by a sum (Theorem 5.2) can be decomposed as a sum of lower-order MHPs.

  • The higher-order derivative ∂A(Hβ)\partial^{A}(H_{\beta}) of an MHP can be expressed in terms of double products and lower-order Hermite polynomials (Theorem 5.4), which relies on Theorem 5.3.

  • There is a formula for the triple product (Theorem 6.2) in terms of double products. For α,β,γ∈N0n\alpha,\beta,\gamma\in\mathbb{N}_{0}^{n} such that ∣γ∣=max⁡(∣α∣,∣β∣,∣γ∣)|\gamma| = \max(|\alpha|,|\beta|,|\gamma|),

⟨Hα,Hβ,Hγ⟩=∑A≤αB≤β∣A∣=∣B∣=k(αA)(βB)⟨HA,HB⟩⟨Hα+β−A−B,Hγ⟩.\langle H_{\alpha},H_{\beta},H_{\gamma}\rangle= \sum_{\substack{A\leq\alpha\\B\leq\beta\\|A|=|B|=k}} \binom{\alpha}{A}\binom{\beta}{B}\langle H_{A},H_{B}\rangle\langle H_{\alpha+\beta-A-B},H_{\gamma}\rangle.

for k:=∣α∣+∣β∣−∣γ∣2k := \frac{|\alpha|+|\beta|-|\gamma|}{2}. Further, the triple product is shown to equal zero when certain conditions on the relative sizes and orders of the multi-indices are satisfied; namely, if

(a) ∣α∣+∣β∣<∣γ∣|\alpha|+|\beta|<|\gamma| or

(b) ∣α∣+∣β∣+∣γ∣|\alpha|+|\beta|+|\gamma| is odd,

then the triple product ⟨Hα,Hβ,Hγ⟩=0\langle H_{\alpha},H_{\beta},H_{\gamma}\rangle= 0.

  • For one-dimensional Hermite polynomials and indicies i,j,k∈N0i,j,k \in\mathbb{N}_{0}, the expression above reduces to the well-known triple product formula (Equation (24)).

The paper is organized as follows. § 2 presents relevant definitions and notational conventions used throughout the document along with necessary background on correlated multivariate Hermite polynomials (MHPs). In § 3, these polynomials serve as the basis functionals for the so-called correlated polynomial chaos expansion (CPCE). § 4 presents an example application of solving a stochastic variant of Burgers’ equation with the CPCE and helps illustrate the practical value of a triple product formula that a computer can evaluate directly. In § 5, we introduce and prove several useful decompositions for MHPs and their derivatives. Then in § 6, we present the paper’s central result: a closed-form expression for the triple product in terms of double products; afterwards, some small examples of the formula’s usage are provided. Subsection 6.1 demonstrates how the expression for the triple product of correlated MHPs reduces to a well-known one-dimensional formula when setting n=1n = 1. Next, § 7 contains a numeric example in which 35+165=20035 + 165 = 200 total9 third-order moments are approximated via simple Monte Carlo, from which we can observe the estimates converging to their predicted formula values as the number of sample realizations becomes sufficiently large. Concluding remarks are provided in section 8.

Definitions, notation, and mathematical preliminaries

Definition 2.1 (Multi-index). An nn-dimensional multi-index α\boldsymbol{\alpha} is an nn-tuple (α1,…,αn)∈N0n(\alpha_1,\ldots,\alpha_n) \in\mathbb{N}_0^n of non-negative integers. Each αi\alpha_i is referred to as the iith element of the multi-index α\boldsymbol{\alpha}.

Unless otherwise specified, a multi-index α\boldsymbol{\alpha} is assumed to have the notation given in Definition 2.1. For multi-indices α,β∈N0n\boldsymbol{\alpha}, \boldsymbol{\beta} \in\mathbb{N}_0^n, one defines [46, 55]

  1. componentwise sum and difference: α±β=(α1±β1,…,αn±βn)\boldsymbol{\alpha} \pm\boldsymbol{\beta} = (\alpha_1 \pm\beta_1,\ldots,\alpha_n \pm\beta_n),

  2. absolute value: ∣α∣:=∥α∥1=α1+⋯+αn|\boldsymbol{\alpha}| := \|\boldsymbol{\alpha}\|_1 = \alpha_1 + \cdots+ \alpha_n, which we call the order of α\boldsymbol{\alpha},

  3. factorial: α!=α1!⋯αn!=∏i=1nαi!\boldsymbol{\alpha}! = \alpha_1! \cdots\alpha_n! = \prod_{i=1}^n \alpha_i!, and the

  4. partial derivative: ∂α=∂1α1⋯∂nαn\partial^{\boldsymbol{\alpha}} = \partial_1^{\alpha_1} \cdots\partial_n^{\alpha_n}.

Moving forward, any α\boldsymbol{\alpha} will denote an nn-dimensional multi-index of order k≥1k \ge1 unless otherwise specified.

Definition 2.2 (Graded lexicographic ordering). For multi-indices α∈N0n\boldsymbol{\alpha} \in\mathbb{N}_0^n, we use a graded lexicographic ordering; that is

α≤β⟺∣α∣<∣β∣ or (∣α∣=∣β∣ and the leftmost digit of α−β is positive).\boldsymbol{\alpha} \le\boldsymbol{\beta} \Longleftrightarrow|\boldsymbol{\alpha}| < |\boldsymbol{\beta}| \text{ or } (|\boldsymbol{\alpha}| = |\boldsymbol{\beta}| \text{ and the leftmost digit of } \boldsymbol{\alpha} - \boldsymbol{\beta} \text{ is positive}).

In all matrices, vectors, and sums indexed by α\boldsymbol{\alpha}, we assume a graded lexicographic ordering (in ascending order) when needed.

For multi-indices, Leibniz’ formula [52] for the generalized product rule is as follows:

∂α(fg)=∑A≤α∂A(f)∂α−A(g).(1)\partial^{\boldsymbol{\alpha}}(fg) = \sum_{\boldsymbol{A} \le\boldsymbol{\alpha}} \partial^{\boldsymbol{A}}(f)\partial^{\boldsymbol{\alpha}-\boldsymbol{A}}(g). \tag*{(1)}

With this notation in place, we present the multi-dimensional analog of the Hermite polynomial.

Correlated multivariate Hermite polynomials

Let (Ω,A,P)(\Omega,\mathcal{A},\mathbb{P}) denote a complete probability space, where

  • Ω\Omega is a nonempty sample space,

  • A\mathcal{A} is a Borel σ\sigma-algebra on Ω\Omega, and

  • P:A→[0,1]\mathbb{P}:\mathcal{A} \to[0,1] is a probability measure.

Let ξ∼N(0,Σ)\boldsymbol{\xi} \sim\mathcal{N}(\boldsymbol{0},\boldsymbol{\Sigma}) be an nn-dimensional random variable with known symmetric positive definite covariance matrix Σ=E(ξξT)\boldsymbol{\Sigma} = \mathbb{E}(\boldsymbol{\xi}\boldsymbol{\xi}^{T}). That is, ξ:(Ω,A)→(Rn,BRn)\boldsymbol{\xi}:(\Omega,\mathcal{A}) \to(\mathbb{R}^n,\mathcal{B}_{\mathbb{R}^n}) where BRn\mathcal{B}_{\mathbb{R}^n} is the Borel σ\sigma-algebra on Rn\mathbb{R}^n. Then ξ\boldsymbol{\xi} has the joint probability density function [39, 46, 63],

φ:Rn→R≥0φ(x;Σ)=1(2π)n/2∣det⁡(Σ)∣1/2exp⁡(−12xTΣ−1x).(2)\varphi:\mathbb{R}^n \to\mathbb{R}_{\ge0} \qquad\varphi(\boldsymbol{x};\boldsymbol{\Sigma}) = \frac{1}{(2\pi)^{n/2}|\det(\boldsymbol{\Sigma})|^{1/2}}\exp\left(-\frac{1}{2}\boldsymbol{x}^{T}\boldsymbol{\Sigma}^{-1}\boldsymbol{x}\right). \tag*{(2)}

The following definition of the multivariate Hermite polynomial is a natural extension of its one-dimensional counterpart [44, 56].

References

  1. [1]A. Alexanderian, On spectral methods for variance based sensitivity analysis, Probability Surveys, 10 (2013), pp. 51 – 68, https://doi.org/10.1214/13-PS219.
  2. [2]J. O. Berger and L. A. Smith, On the statistical formalism of uncertainty quantification, Annual Review of Statistics and Its Application, 6 (2019), pp. 433–460, https://doi.org/10.1146/annurev-statistics-030718-105232.
  3. [3]M. Brani̇cki and A. J. Majda, Fundamental limitations of polynomial chaos for uncertainty quantification in systems with intermittent instabilities, Communications in Mathematical Sciences, 11 (2013), pp. 55–103, https://api.semanticscholar.org/CorpusID:16416936.DOI
  4. [4]P. Constantine, A primer on stochastic Galerkin methods, Feb. 2007, https://www.cs.colorado.edu/~paco3637/docs/constantine2007primer.pdf (accessed 2020-02-06).
  5. [5]P. Constantine, Spectral methods for parametrized matrix equations, PhD thesis, Stanford University, 2009.
  6. [6]B. Debusschere, Intrusive Polynomial Chaos Methods for Forward Uncertainty Propagation, Springer International Publishing, Cham, 2017, pp. 617 – 636, https://doi.org/10.1007/978-3-319-12385-1_19.
  7. [7]B. Debusschere, H. N. Najm, P. Pèbay, O. M. Knio, R. Ghanem, and O. P. L. Maître, Numerical challenges in the use of polynomial chaos representations for stochastic processes, SIAM Journal on Scientific Computing, 26 (2004), pp. 698 – 719, https://doi.org/10.1137/S1064827503427741.
  8. [8]P. Dellaportas and G. O. Roberts, An Introduction to MCMC, Springer New York, New York, NY, 2003, pp. 1–41, https://doi.org/10.1007/978-0-387-21811-3_1.
  9. [9]A. Dembow, Probability theory: STAT310/MATh230, 2019, https://web.stanford.edu/class/stats310a/notes.pdf (accessed 2025-07-10).
  10. [10]C. F. Dunkl and Y. Xu, Orthogonal Polynomials of Several Variables, Encyclopedia of Mathematics and its Applications, Cambridge University Press, 2nd ed., 2014, https://doi.org/10.1017/CBO9781107786134.
  11. [11]R. Durrett, Probability: Theory and Examples, Cambridge Series in Statistical and Probabilistic Mathematics, Cambridge University Press, 3rd ed., 2005.
  12. [12]M. Eldred, C. Webster, and P. G. Constantine, Evaluation of non-intrusive approaches for Wiener-Askey generalized polynomial chaos, AIAA 2008-1892, (2008), https://doi.org/10.2514/6.2008-1892.
  13. [13]O. G. Ernst, A. Mugler, H. J. Starkloff, and E. Ullmann, On the convergence of generalized polynomial chaos expansions, ESAIM: M2AN, 46 (2012), pp. 317 – 339, https://doi.org/10.1051/m2an/2011045.
  14. [14]J. Feinberg, V. E. Eck, and H. P. Langtangen, Multivariate polynomial chaos expansions with dependent variables, SIAM Journal on Scientific Computing, 40 (2018), pp. A199 – A223, https://doi.org/10.1137/15M1020447.
  15. [15]G. Fishman, Monte Carlo: Concepts, Algorithms, and Applications, Springer Series in Operations Research and Financial Engineering, Springer, 2nd ed., 1996, https://doi.org/10.1007/978-1-4757-2553-7.
  16. [16]T. Gerstner and M. Griebel, Numerical integration using sparse grids, Numerical Algorithms, 18 (1998), pp. 209–232, https://doi-org.stanford.idm.oclc.org/10.1023/A:1019129717644.DOI
  17. [17]R. Ghanem, D. Higdon, and H. Owhadi, Introduction to Uncertainty Quantification, Springer International Publishing, Cham, 2017, pp. 3–6, https://doi.org/10.1007/978-3-319-12385-1_1.
  18. [18]R. G. Ghanem and P. D. Spanos, Stochastic Finite Elements: A Spectral Approach, Springer-Verlag, Berlin, Heidelberg, 1991.DOI
  19. [19]G. H. Golub and C. F. V. Loan, Matrix Computations, The Johns Hopkins University Press, 4th ed., 2013.
  20. [20]D. Gottlieb and C. W. Shu, On the Gibbs phenomenon and its resolution, SIAM Review, 39 (1997), pp. 644 – 668, http://www.jstor.org/stable/2132695.
  21. [21]M. Griebel, Sparse Grids for Higher Dimensional Problems, London Mathematical Society Lecture Note Series, Cambridge University Press, 2006, pp. 106–161, https://doi.org/10.1017/CBO9780511721571.004.
  22. [22]M. Gunzburger, C. G. Webster, and G. Zhang, Sparse Collocation Methods for Stochastic Interpolation and Quadrature, Springer International Publishing, Cham, 2017, pp. 717–762, https://doi.org/10.1007/978-3-319-12385-1_29.
  23. [23]E. Hewitt and R. E. Hewitt, The Gibbs-Wilbraham phenomenon: An episode in Fourier analysis, Arch. Hist. Exact Sci., 21 (1979), pp. 129 – 160, https://doi.org/10.1007/ BF00330404.
  24. [24]W. HUBERTS, W. P. DONDERs, T. DELHAAS, AND F. N. VAN DE VOSSE, Applicability of the polynomial chaos expansion method for personalization of a cardiovascular pulse wave propagation model, International Journal for Numerical Methods in Biomedical Engineering, 30 (2014), pp. 1679 – 1704, https://doi.org/10.1002/cnm.2695.
  25. [25]J. JAKEMAN ET AL., Polynomial chaos expansions for dependent random variables, Computer Methods in Applied Mechanics and Engineering, 351 (2019), pp. 643–666, https://doi.org/10.1016/j.cma.2019.03.049.
  26. [26]G. E. KARNIADAKIS ET AL., Generalized polynomial chaos solution for differential equations with random inputs., tech. report, 2005-1, Seminar for Applied Mathematics, ETH Zürich, Zürich, Switzerland, Jan 2005.
  27. [27]K. KWANG-KI AND R. BRAATZ, Generalised polynomial chaos expansion approaches to approximate stochastic model predictive control, International Journal of Control, 86 (2013), pp. 1324–1337, https://doi.org/10.1080/00207179.2013.801082.
  28. [28]H. LI AND D. ZHANG, Probabilistic collocation method for flow in porous media: Comparisons with other stochastic methods, Water Resources Research, 43 (2007), https://doi.org/10.1029/2006WR005673.
  29. [29]J. LIU, Monte Carlo Strategies in Scientific Computing, Springer, 2001.
  30. [30]W. K. LIU, T. BELYSCHKO, AND A. MANI, Random field finite elements, International Journal for Numerical Methods in Engineering, 23 (1986), pp. 1831–1845, https://onlinelibrary.wiley.com/doi/abs/10.1002/nme.1620230104.DOI
  31. [31]J. L. LUMLEY, The structure of inhomogeneous turbulent flows, Atmospheric Turbulence and Radio Wave Propagation, (1967), https://cir.nii.ac.jp/crid/157198075051475712.
  32. [32]L. LYMAN AND G. IACCARINO, A bluff-and-fix algorithm for polynomial chaos methods, in Computational Science – ICCS 2020, V. V. Krzhizhanovskaya, G. Závodszky, M. H. Lees, J. J. Dongarra, P. M. A. Sloot, S. Brissos, and J. Teixeira, eds., Cham, 2020, Springer International Publishing, pp. 742 – 756, https://doi.org/10.1007/978-3-030-50436-6_55.
  33. [33]L. LYMAN AND G. IACCARINO, Second order moments of multivariate Hermite polynomials in correlated random variables, in Computational Science – ICCS 2021, M. Pazynski, D. Kranzlmüller, V. V. Krzhizhanovskaya, J. J. Dongarra, and P. M. A. Sloot, eds., Cham, 2021, Springer International Publishing, pp. 698–712, https://doi.org/10.1007/978-3-030-77980-1_53.
  34. [34]O. L. MAÎTRE AND O. KNIO, Spectral Methods for Uncertainty Quantification: With Applications to Computational Fluid Dynamics, January 2010, https://doi.org/10.1007/978-90-481-3520-2.
  35. [35]O. L. MAÎTRE, O. KNIO, H. N. NAJM, AND R. G. GHANEM, Uncertainty propagation using Wiener–Haar expansions, Journal of Computational Physics, 197 (2004), pp. 28–57, https://www.sciencedirect.com/science/article/pii/S0021999103006181.
  36. [36]H. MATTHIES AND A. KEESE, Galerkin methods for linear and nonlinear elliptic stochastic partial differential equations, Computer Methods in Applied Mechanics and Engineering, 194 (2005), pp. 1295 – 1331, https://doi.org/10.1016/j.cma.2004.05.027.
  37. [37]A. NATAF, Détermination des distributions de probabilités dont les marges sont données, C. R. Acad. Sci. Paris, 255 (1962), pp. 42 – 43.
  38. [38]M. NAVARRO, J. WITTEVEEN, AND J. BLOM, Polynomial chaos expansion for general multivariate distributions with correlated variables, 2014, https://doi.org/10.48550/ARXIV.1406.5483.
  39. [39]S. NOREDDINE AND I. NOURDIN, On the Gaussian approximation of vector-valued multiple integrals, Journal of Multivariate Analysis, 102 (2011), pp. 1008 – 1017, https://doi.org/10.1016/j.jmva.2011.02.001.
  40. [40]W. OBERKAMPF AND T. TRUCANO, Validation methodology in computational fluid dynamics, AIAA 2000-2549, (2000), https://doi.org/10.2514/6.2000-2549.
  41. [41]A. B. OWEN, Monte Carlo theory, methods and examples, https://artowen.su.domains/mc/, 2013.
  42. [42]A. B. OWEN, Practical Quasi-Monte Carlo Integration, https://artowen.su.domains/mc/practicalqmc.pdf, 2023.
  43. [43]Z. PERKÓ, S. R. VAN DER VOORT, ET AL., Fast and accurate sensitivity analysis of impt treatment plans using polynomial chaos expansion, Physics in Medicine & Biology, 61 (2016), p. 4646, https://doi.org/10.1088/0031-9155/61/12/4646.
  44. [44]M. P. PETTERSSON, G. IACCARINO, AND J. NORDSTRÖM, Polynomial Chaos Methods for Hyperbolic Partial Differential Equations: Numerical Techniques for Fluid Dynamics Problems in the Presence of Uncertainties, Springer International, Switzerland, 2015, https://doi.org/10.1007/978-3-319-10714-1.
  45. [45]P. Pettersson, G. Iaccarino, and J. Nordström, A stochastic Galerkin method for the Euler equations with Roe variable transformation, Journal of Computational Physics, 257 (2014), pp. 481 – 500, https://www.sciencedirect.com/science/article/pii/S0021999113006797.DOI
  46. [46]S. Rahman, Wiener–Hermite polynomial expansion for multivariate Gaussian probability measures, Journal of Mathematical Analysis and Applications, 454 (2017), pp. 303 – 334, http://www.sciencedirect.com/science/article/pii/S0022247X17304250.arxiv.org/abs/1704.07912
  47. [47]Z. Rosario, Precision margin: First-principles margins for aircraft design under uncertainty, PhD thesis, Stanford University, 2020.
  48. [48]M. Rosenblatt, Remarks on a Multivariate Transformation, The Annals of Mathematical Statistics, 23 (1952), pp. 470 – 472, https://doi.org/10.1214/aoms/1177729394.
  49. [49]W. Rudin, Principles of Mathematical Analysis, McGraw-Hill Book Company, Inc., New York-Toronto-London, 1953.
  50. [50]J. Salazar, The Polynomial Chaos Method with Applications to Random Differential Equations, PhD thesis, The University of Texas at Arlington, 2013.
  51. [51]J. Salverda, Probabilistic treatment planning with polynomial chaos expansion for proton therapy, master’s thesis, Delft University of Technology, 2019.
  52. [52]A. Sasane, A Friendly Approach to Functional Analysis, World Scientific Publishing Company, 2007, ch. 6.4, pp. 248 – 249, https://books.google.com/books/about/A_Friendly_Approach_To_Functional_Analysis.html?id=vEQyDwAAQBAJ
  53. [53]M. Schober, S. Särkkä, and P. Hennig, A probabilistic model for the numerical solution of initial value problems, 2017, https://arxiv.org/abs/1610.05261.
  54. [54]K. K. Sepahvand, S. Marburg, and H.-J. Hardtke, Uncertainty quantification in stochastic systems using polynomial chaos expansion, International Journal of Applied Mechanics, 02 (2012), https://doi.org/10.1142/S1758825110000524.
  55. [55]D. Slepian, On the symmetrized Kronecker power of a matrix and extensions of Mehler’s formula for Hermite polynomials, SIAM Journal on Mathematical Analysis, 3 (1972), pp. 606 – 616, https://doi.org/10.1137/0503060.
  56. [56]R. Smith, Uncertainty Quantification: Theory, Implementation, and Applications, SIAM - Society for Industrial and Applied Mathematics, 2013.DOI
  57. [57]P. Sochala and O. P. Le Maître, Polynomial chaos expansion for subsurface flows with uncertain soil parameters, Advances in Water Resources, 62 (2013), pp. 139–154, https://www.sciencedirect.com/science/article/pii/S030917081300184X.DOI
  58. [58]C. Soize and R. Ghanem, Physical systems with random uncertainties: Chaos representations with arbitrary probability measure, SIAM Journal on Scientific Computing, 26 (2004), pp. 395 – 410, https://doi.org/10.1137/S1064827503424505.
  59. [59]R. P. Stanley, Enumerative Combinatorics, Vol. 2, Cambridge University Press, 2001.
  60. [60]G. Szegő, Orthogonal Polynomials, American Math. Soc: Colloquium publ, American Mathematical Society, 1975, https://books.google.com/books?id=ZOhm sXlcY0C.
  61. [61]X. Wang et al., A data-driven sparse polynomial chaos expansion method to assess probabilistic total transfer capability for power systems with renewables, IEEE Transactions on Power Systems, 36 (2021), pp. 2573–2583, https://doi.org/10.1109/TPWRS.2020.3034520.
  62. [62]N. Wiener, The homogeneous chaos, American Journal of Mathematics, 60 (1938), pp. 897 – 936, http://www.jstor.org/stable/2371268.DOI
  63. [63]D. Xiu, Generalized (Wiener-Askey) Polynomial Chaos, PhD thesis, Brown University, 2004.
  64. [64]D. Xiu, Stochastic Collocation Methods: A Survey, Springer International Publishing, Cham, 2017, pp. 699–716, https://doi.org/10.1007/978-3-319-12385-1_26.
  65. [65]D. Xiu et al., Modeling uncertainty in steady state diffusion problems via generalized polynomial chaos, Computer Methods in Applied Mechanics and Engineering, 191 (2002), pp. 4927–4948, https://www.sciencedirect.com/science/article/pii/S0045782502004218.
  66. [66]D. Xiu and G. E. Karniadakis, The Wiener–Askey polynomial chaos for stochastic differential equations, SIAM J. Sci. Comput., 24 (2002), pp. 619 – 644, https://doi.org/10.1137/S1064827501387826.

Paper details

vlnc:W_6gn5fj65
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

None listed by the authors

View statement
Looking for related papers…
Contents