General Introduction to Polynomial

Chaos and Collocation Methods

Chris Lacor and Éric Savin

Polynomial Chaos Methodology

The polynomial chaos methodology (PCM) is a rather recent approach, which offers
a large potential for computational fluid dynamics (CFD) related non-deterministic
simulations, as it allows the treatment of a large variety of stochastic variables and
properties that can be described by probability density functions (PDFs). The method
is based on a spectral representation of the uncertainty where the basis polynomials
contain the randomness, described by random variables � with values in a set � ,
and the unknown expansion coefficients are deterministic, resulting in deterministic
equations. More specifically, if u is a random variable indexed by a spatial variable
� ∈ D ⊆ ℝd (typically, d = 3 in physical space) and time t ≥ 0, the so-called poly-
nomial chaos expansion (PCE) reads:

P
∑
P
u(�, t, � ) ≃ ℙ [u](�, t, � ) = ui (�, t)�i (�� ) . (1)
i=0

In the above, ui are the deterministic unknown expansion coefficients and represent
the random mode i of the random variable u. �i are N-variate polynomials which
are functions of � = (�1 , �2 , … , �N ) where �j is a random variable with values in a
set �j . N is the number of input uncertainties which is also the number of random

C. Lacor (✉)
Mechanical Engineering Department, Vrije Universiteit Brussel,
Pleinlaan 2, 1050 Brussels, Belgium
e-mail: [email protected]
É. Savin
Computational Fluid Dynamics Department, Onera–The French Aerospace Lab,
29 avenue de la Division Leclerc, 92322 Châtillon, France
e-mail: [email protected]

© Springer International Publishing AG, part of Springer Nature 2019 109

C. Hirsch et al. (eds.), Uncertainty Management for Robust Industrial Design
in Aeronautics, Notes on Numerical Fluid Mechanics and Multidisciplinary
Design 140, https://doi.org/10.1007/978-3-319-77767-2_7
110 C. Lacor and Savin

dimensions. It is assumed that these variables are independent and are real valued,
and hence � = �1 × �2 × ⋯ × �N ⊆ ℝN . Input uncertainties could, e.g., be associ-
ated with uncertain operational conditions or uncertainty in the geometry. For an
external flow around an airplane, the inlet Mach number, angle of attack, inlet pres-
sure, etc., are examples of operational conditions. Geometrical uncertainties are then
uncertainties on the shape of the shape of the plane due to manufacturing tolerances.
It is clear that because of the uncertain input, any flow variable, say u becomes also
uncertain, and can therefore be described as in Eq. (1). The total number of terms
P + 1 used in (1) depends on the highest order of the polynomial that is used (denoted
by p) and on the number of random dimensions. One has, see [1]:

(N + p)!
P+1= . (2)
N!p!

The methodology was originally formulated by Wiener [2] and was much later redis-
covered and used for CFD applications by several groups, e.g., Xiu and Karniadakis
[3], Lucor et al. [4], Le Maître et al. [5], Mathelin et al. [6], and Walters and Huyse
[7] among others.
In the original method of Wiener [2], the projection basis �i is constituted by
Hermite polynomials. These are optimal for random variables with Gaussian dis-
tribution. Optimal means that, for increasing polynomial order, the expansion will
quickly converge in the mean-square sense. The condition for optimality is that the
polynomials are orthogonal with a weighting function � ↦ WN (�� ) which is exactly
the PDF of the set of random variables, i.e.:

∫�
�i (�� )�j (�� )WN (�� )d�� ∶= ⟨�i , �j ⟩ = �j �ij , (3)

where �ij is the Kronecker symbol, and �j is a normalization constant. With a

proper scaling though, one can always normalize the polynomial basis such that
�j ∶= ⟨�j , �j ⟩ = 1 ∀j. In the case of a multivariate Gaussian distribution, the Her-
mite polynomials satisfy the condition above with WN given by:

WN (�� ) ≡ √
1 1
exp(− � ⋅ � ) , (4)
(2�)N 2

∑N
where � ⋅ � = j=1 �j2 is the standard Euclidian scalar product in ℝN . Note that
because of the independence of the uncertainties, the PDF is the product of the PDF
∏
of each of the uncertainties, i.e., WN (�� ) = Nj=1 W1 (�j ) as defined above for Gaussian
uncertainties.
For uncertainties with other distributions, the orthogonality condition (3) gives
adapted polynomials, see e.g. [3], leading to the so-called Askey scheme; for exam-
ple, as already mentioned Hermite polynomials for Gaussian distributions, and
further Charlier polynomials for Poisson distributions, Laguerre polynomials for
Gamma distributions, Jacobi polynomials for Beta distributions, etc. In case of less
General Introduction to Polynomial Chaos and Collocation Methods 111

common distributions, an optimal PCM can always be found by constructing the

polynomials via a Gram-Schmidt procedure; see Witteveen and Bijl [8], in order to
satisfy (3). It should be noted that, if the optimal polynomials are not used, the PCM
will also converge (with increasing order) in the mean-square sense but much slower
than the exponential convergence with optimal polynomials; see [1].
In cases where the response of the system shows a localized sharp variation or
a discontinuous change, local expansions may be more efficient than expansions
with global polynomials, whose convergence will deteriorate due to the Gibbs phe-
nomenon. This has led to developments using wavelet expansions [9] and to multi-
element polynomial chaos [10, 11]. In the latter case, the random space is subdi-
vided in smaller elements in which new random variables are defined with associated
orthogonal polynomials that are constructed numerically.
As already mentioned, the dimension of the problem N is determined by the num-
ber of independent random input variables. In case of a random process (as opposed
to a random variable), a Karhunen-Loève expansion (also known as Principal Com-
ponent Analysis or Proper Orthogonal Decomposition) [12, 13] can be applied to the
correlation function R(�, �) of the random process u(�) indexed by � ∈ D, to decom-
pose the random input process in a set of uncorrelated random variables. Assuming
∫D R(�, �)d� < +∞ (which is untrue for a stationary process with D ≡ ℝd ) and solv-
ing the eigenvalue problem:

∫D
R(�, �)�i (�)d� = �i �i (�) (5)

with �i (�) the eigenfunctions and �i the eigenvalues, the Karhunen-Loève expansion
of the random field u(�) becomes:
∑√
u(�) − u(�) = �i �i �i (�) , (6)
i

where the �i s are uncorrelated random variables, and u(�) is the mean value at the
indexation point �. Note that if the process u is Gaussian, the random variables �i are
Gaussian as well, and hence, they are mutually independent.
A geometrical uncertainty is typically a random process where the coordinates of
a geometry are uncertain with some specific correlation length. Depending on the
correlation length of the process, the eigenvalues �i become quickly very small, so
that only few terms in the summation above have to be kept. This is not the case
however for a very short correlation length (e.g., white noise) resulting in a high-
dimensional chaos expansion for such processes. Non-Gaussian random processes
are much more difficult to treat than Gaussian [14]. In the former case, mean and
covariance are far from sufficient to completely specify the process. This remains an
active area of research.
The PCM can be implemented either in an intrusive or in a non-intrusive way
as follows.
112 C. Lacor and Savin

Intrusive Polynomial Chaos

In an intrusive PCM, the polynomial expansion of the unknown variables, Eq. (1),
is introduced in the model, e.g., for CFD applications, the Navier–Stokes equations.
Each unknown u is therefore replaced with its expansion coefficients ui . The num-
ber of unknowns is therefore basically multiplied with a factor P + 1, which can be
quite high for high stochastic dimensions and/or high polynomial order. In addition,
the model, e.g., CFD code, has to be adapted. The required effort for extending a
deterministic CFD code with the intrusive PCM depends on the characteristics of
the code: computer language, structured/unstructured, handling of data storage, etc.
In the framework of the NODESIM-CFD EU project, an intrusive PCM was imple-
mented in the commercial code Fine/Turbo of NUMECA. This has led to one of
the first applications of intrusive PCM to three-dimensional turbulent Navier–Stokes
flows [15]. The number of additional lines of code is very limited, compared to the
length of the original, deterministic code. However, changes are not restricted to a
local part of the code. This increases the risk of introducing bugs and requires some-
one who is very familiar with all aspects of the code. This is a big disadvantage
compared to non-intrusive PC and the main reason why the application of intrusive
PCM in commercial codes is very limited.
Nonetheless intrusive methods are more flexible and in general more precise than
non-intrusive methods; see Aleksev et al. [16]. This is also confirmed by Xiu [14],
who mentions that the intrusive method offers the most accurate solutions involving
the least number of equations in multi-dimensional random spaces, even though the
resulting equations are coupled.
It is to be noted that the treatment of geometrical uncertainties needs a different
approach compared to operational uncertainties. A possibility is to use a transforma-
tion such that the deterministic problem in a stochastic domain becomes a stochas-
tic problem in a deterministic domain, e.g., Xiu and Tartakovsky [17]. An alterna-
tive is the use of a so-called fictitious domain method [18, 19], or by introducing
the uncertainty directly in the surface normals within a control volume approach
[20, 21].

Non-intrusive Polynomial Chaos

In the UMRIDA EU project, all PCM contributions relate to non-intrusive

approaches. Basically, two different classes of approaches have been formulated:
(i) the so-called projection method, which is based on a numerical evaluation of
the Galerkin integrals; see Le Maître et al. [5, 22, 23] and Nobile et al. [24]; (ii)
regression methods based on a selected set of sample points; see Berveiller et al.
[25], and Hosder et al. [26–28].
General Introduction to Polynomial Chaos and Collocation Methods 113

In the projection methods, starting from Eq. (1), the projection on �j yields:

P
∑
∫� ∫�
u(�, t, � )�j (�� )WN (�� )d�� = ui (�, t) �i (�� )�j (�� )WN (�� )d��
i=0 (7)
j
= �j u (�, t)

The last equation results from the orthogonality condition (3) and can be considered
as an equation for the unknown expansion coefficient uj . It requires the evaluation
of the integral in the left-hand side. A numerical quadrature formula is used. For a
single variable parameter, it reads:
q
∑
∫�1
u(�, t, �)�j (�)W1 (�)d� ≃ wl u(�, t, � l )�j (� l ) . (8)
l=1

The evaluation of the sum in the right-hand side requires an evaluation of the
unknown u in q sample points {� l }1≤l≤q in �1 associated to q weights {wl }1≤l≤q .
Depending on the weighting function (PDF) W1 , adapted Gaussian quadrature for-
mulations exist for an accurate evaluation: With q sample points, a polynomial
of order 2q − 1 is integrated exactly in one dimension. Examples are the Gauss-
Legendre quadrature (W1 = 1∕2 corresponding to a uniform distribution), the Gauss-
Hermite quadrature (W1 given by Eq. (4) in one dimension), etc. For a PCM of order
p, one takes q = p + 1. This guarantees exact quadrature if u(�, t, �) can be described
by a polynomial of maximum order p + 1.
This extends to multiple stochastic dimensions by using a full-tensor product
quadrature with Q = qN sample points. This approach quickly becomes very expen-
sive for high-order and high stochastic dimensions. This has led to the use of sparse
grid sampling techniques, avoiding the full-tensorial sampling, e.g., the Smolyak
scheme [29]. Sparse grid schemes can be combined with the non-nested Gaus-
sian quadratures invoked above, as well as with nested quadratures, e.g., Clenshaw-
Curtis, Gauss-Patterson [30–33]. More recently, adaptive algorithms have been
developed that further reduce the cost [34–36]. The choice of quadrature sets is
discussed further on in section “Choices of Interpolation Set” in relation with the
stochastic collocation method. Alternatively, the numerical quadrature can also be
achieved using Monte Carlo simulation [37, 38], or Latin Hypercube sampling [39].
All in all, the evaluation of the left-hand side of Eq. (7) using Q sampling points
{�� l }1≤l≤Q in � associated to Q weights {wl }1≤l≤Q yields:

P
( Q
)
∑ 1 ∑ l
u(�, t, � ) ≃ ℙPQ [u](�, t, � ) = l l
w u(�, t, � )�i (�� ) �i (�� ) . (9)
i=0
� i l=1

In linear regression methods, the stochastic problem is solved in S samples in

stochastic space. For each sample s, Eq. (1) can be written as:
114 C. Lacor and Savin

P
∑
u(�, t, � s ) = ui (�, t)�i (�� s ) . (10)
i=0

This leads to S equations for the P + 1 unknowns ui . Note that this forms a linear
system. In order to make the solution less dependent on the choice of the samples,
oversampling is used and the system is solved with regression (i.e., the least squares
method); see Berveiller et al. [25] and Hosder et al. [26–28]. As a rule of thumb,
S = 2(P + 1) is a good choice; see [27]. Different sampling techniques can be used
such as Random, Latin Hypercube, Hammersley [27], roots of Hermite polynomials
of order p + 1 (for PCM of order p with Gaussian uncertainties) [25], Sobol’ quasi-
random sampling [40], etc.
In case of geometrical uncertainties, each of the different samples–both in the pro-
jection and the regression method–will correspond to a different geometry. Geomet-
rical uncertainties therefore require no special treatment in contrast with the intrusive
method.

The Collocation Method

The stochastic collocation (SC) method based on Lagrange interpolation has been
introduced in [41] and developed further on in e.g. [24, 42–45]. Examples of appli-
cations can be found in [46–51] among others. Along the same lines as Eq. (1),
the SC expansion is formed as a sum of multi-dimensional Lagrange interpolation
polynomials with respect to the N–dimensional random input variable � . Lagrange
polynomials interpolate a set of points in one dimension {�1l }1≤l≤q1 in a bounded
interval �1 by the following functional form:
q1
∏ � − �1k
Ll (�) = , (11)
�1l − �1k
k≠l
k=1

such that Ll (�1k ) = �kl , 1 ≤ k, l ≤ q1 ; in addition, all Ll ’s have order q1 − 1. For inter-
polation in multiple dimensions, the tensor product of one-dimensional Lagrange
polynomials can be formed. Eventually at this stage, it is assumed that the interpola-
tion set is formed by tensorization of one-dimensional sets. In other words, structured
interpolation sets are considered, for multivariate Lagrange interpolation on unstruc-
tured, arbitrary sets of nodes still raises numerous theoretical and practical difficul-
ties. Letting � = (l1 , l2 … lN ) be a multi-index in ℕN ⧵ {�}, the multi-dimensional
Lagrange polynomial L� reads:

L� (�� ) = Ll1 (�1 ) ⊗ Ll2 (�2 ) ⊗ ⋯ ⊗ LlN (�N ) , (12)

where different interpolation sets {�jl }1≤l≤qj in different intervals �j may possibly be
used for each different dimension j. If Q is now the total number of such multi-
General Introduction to Polynomial Chaos and Collocation Methods 115

dimensional interpolation points counted by a single index l, {�� l }1≤l≤Q , the SC

expansion of the random field u reads:

Q
∑
u(�, t, � ) ≃ �Q [u](�, t, � ) = u(�, t, � l )Ll (�� ) , (13)
l=1

where the expansion coefficients are the random field evaluated at � l .

Choices of Interpolation Set

The key issue of the SC method is the choice of appropriate interpolation sets. A
natural, straightforward choice is quadrature nodes and weights as in Eq. (8). Multi-
dimensional quadrature sets � (N, Q) = {�� l , wl }1≤l≤Q , where � l is the l-th node in
∏N
� = j=1 �j and wl is the corresponding weight, may be constructed from one-
dimensional (univariate) quadrature sets by full tensorization or sparse tensorization,
using Smolyak’s algorithm [29] as already invoked above.
Univariate Gauss quadratures �(1, q1 ) based on q1 integration points are tailored
to integrate on �1 ≡ [a, b] a smooth function � ↦ f (�):
q1 −r r
∑ ∑
∫�1
f (�)W1 (�)d� ≃ wl f (� l ) + wq1 −r+m f (� q1 −r+m ) , (14)
l=1 m=1

such that this rule turns to be exact for univariate polynomials up to the order
2q1 − 1 − r. Here, r is the number of fixed nodes of the rule, typically the bounds
a, b. Depending on the choice of r, different terminologies are used:
∙ r = 0 is the classical Gauss rule;
∙ r = 1 is the Gauss-Radau (GR) rule, choosing � q1 = a or � q1 = b for instance;
∙ r = 2 is the Gauss-Lobatto (GL) rule, choosing � q1 −1 = a and � q1 = b for instance.
Multivariate quadratures may subsequently be obtained by full or sparse tensoriza-
tion of these one-dimensional rules. Firstly, a fully tensorized grid is obtained by the
straightforward product rule:

N
⨂
� (N, Q) = �(1, qj ) , (15)
j=1

∏N
which contains Q = j=1 qj grid points in � . Secondly, a sparse quadrature rule
can be derived thank to the Smolyak algorithm [29]. The so-called k–th level, N-
dimensional Smolyak sparse grid �̂ (N, k) is obtained by the following linear combi-
nation of product formulas [52]:
116 C. Lacor and Savin

Fig. 1 Two-dimensional (N = 2) nodes based on a non-nested, one-dimensional Gauss-Lobatto

quadrature rule with q = 9 nodes. Left: fully tensorized grid (Q = 81). Right: sparse tensorized
grid from Smolyak’s algorithm with k = 9 (Q = 193)

k−1
∑ ∑
̂ (N, k) =
� �(1, q1 ) ⊗ ⋯ ⊗ �(1, qN ) . (16)
l=k−N q1 +⋯+qN =N+l

Clearly, the above sparse grid is a subset of the full-tensor product grids. It typically
contains Q ∼ (2N)k−1 ∕k − 1! nodes in � whenever N ≫ 1 and k is fixed. By a direct
extension of the arguments divised in [31, 33], it can be shown that provided the
univariate quadrature rules �(1, q) are exact for all univariate polynomials of order
up to 2q − 1 (Gauss rules) or 2q − 3 (GL rules), the foregoing rule is exact for all
N–variate polynomials of total order up to 2k − 1 or 2k − 3, respectively. Figure 1
displays for example the two-dimensional full and sparse rules for an underlying
univariate GL quadrature (14) with q = 9 nodes and W1 (�) = (1 − � 2 )3 , �1 = [−1, 1].
For this example:

̂ (2, 9) = �(1, 2) ⊗ �(1, 7) + �(1, 3) ⊗ �(1, 6) + �(1, 4) ⊗ �(1, 5)

�
+ �(1, 2) ⊗ �(1, 8) + �(1, 3) ⊗ �(1, 7) + �(1, 4) ⊗ �(1, 6)
+ �(1, 5) ⊗ �(1, 5) + perm.

Here, Q = 193, compared to Q = 81 with the fully tensorized rule (15). In [53], it has
been observed that sparse quadratures outperform fully tensorized quadratures with
non-nested underlying one-dimensional rules whenever N ≥ 4, though. If �(1, qi ) is
now Clenshaw-Curtis (CC) univariate quadrature of i-th level for i > 1, such that:

1 ≤ l ≤ qi = 2i−1 + 1 ,
(l − 1)�
� l = − cos ,
qi − 1

then the associated third-level bivariate sparse rule as constructed in, e.g., [32] for,
say, q = 9 is:
General Introduction to Polynomial Chaos and Collocation Methods 117

Fig. 2 Two-dimensional (N = 2) nodes based on a nested, one-dimensional Clenshaw-Curtis

quadrature rule with q = 9 nodes. Left: fully tensorized grid (Q = 81). Right: sparse tensorized
grid from Smolyak’s algorithm with k = 3 (Q = 29)

̂ (2, 3) = �(1, 1) ⊗ �(1, 5) + �(1, 3) ⊗ �(1, 3)

�
(17)
+ �(1, 1) ⊗ �(1, 9) + �(1, 3) ⊗ �(1, 5) + perm.

The underlying univariate CC rules �(1, qi ) are nested, that is, �(1, qi ) ⊂ �(1, qi+1 ),
and consequently, the multivariate rules are nested as well, � ̂ (N, k) ⊂ �̂ (N, k + 1).
They are in addition exact at least for all multivariate polynomials of total order
k [32]. Figure 2 displays the two-dimensional full rule (15) and third-level sparse
rule (17) corresponding to the univariate CC quadrature with q = 9 nodes. The total
number of nodes is significantly reduced with such a nested rule.

Link with Polynomial Chaos

The multi-dimensional Lagrange polynomials may be expanded on the

multi-dimensional polynomial chaos basis {�i }0≤i≤P as in Eq. (1):

1 ≤ l ≤ Q,
P
∑
Ll (�� ) = ⟨Ll , �i ⟩�i (�� ) ,
i=0
∑
where P is given by Eq. (2) with polynomial total order p = Nj=1 qj − N. The
expansion coefficients �li ∶= ⟨Ll , �i ⟩ can be evaluated with the quadrature rule
{�� l , wl }1≤l≤Q also used as the interpolation set:

Q
1 ∑ 1
�li ≃ wm Ll (�� m )�i (�� m ) = wl �i (�� l ) ,
�iQ m=1 �iQ

where the second equality stems from the very definition of Lagrange polynomials.
∑Q
Here, �iQ = l=1 wl (�i (�� l ))2 is the normalization constant for the polynomial chaos,
118 C. Lacor and Savin

which is simply �iQ = �i if the quadrature rule integrates exactly polynomials of total
order 2p. Consequently, the SC expansion (13) of the random field u reads:

Q P
∑ ∑ 1
� [u](�, t, � ) ≃ �Q
Q
P
[u](�, t, � ) = u(�, t, � l ) Q
wl �i (�� l )�i (�� )
l=1 i=0 �i
( ) (18)
P Q
∑ 1 ∑ l l
l
= w u(�, t, � )�i (�� ) �i (�� ) .
i=0 �iQ l=1

The bracketed sum above is the evaluation of the PC expansion coefficients ui by

the quadrature rule at hand. Hence, both PC and SC expansions are mathematically
equivalent, �Q
P
≡ ℙPQ , though they are numerically slightly different [54].

Application to Uncertainty Quantification (UQ)

Once the polynomial expansion (1) or (13) has been derived, the first moments and/or
cumulants of the random field u can be computed using a quadrature rule � (N, Q)
and associated evaluations u(�, t, � l ), 1 ≤ l ≤ Q. Indeed, for a regular function u ↦
f (u), one can estimate a mean output functional by:

Q
∑
∫�
�{f (u)}(�, t) = f (u(�, t, � ))WN (�� )d�� ≃ wl f (u(�, t, � l )) .
l=1

The mean � is obtained for f (u) = u, the variance � 2 is obtained for f (u) = (u − �)2 ,
the skewness �1 for f (u) = ( u−�
�
)3 , the kurtosis �2 for f (u) = ( u−�
�
)4 , etc. More gen-
erally, the j-th moment mj is obtained for f (u) = u and may be used to compute the
j

characteristic function �U :
+∞
∑ mj
∫
�U (V) = eiU⋅V WU (dU) = (iU)j ,
j=0
j!

where by the causality principle (or transport of PDFs) for the random variable U ∼
u(⋅, � ) one has:
| du−1 |
WU (dU) = || | WN (u−1 (dU)) .
|
| dU |
Sobol’ sensitivity indices or global sensitivity indices may be computed alike; see
[14, 53, 55–57] and references therein. Denoting by Ij the set of indices corre-
sponding to the polynomials �k depending only on the j-th variable parameter �j , the
main-effect PCE-based Sobol’ indices are given by (see e.g. Sudret [57]):
General Introduction to Polynomial Chaos and Collocation Methods 119

1 ∑
Sj (�, t) = � (uk (�, t))2 ,
� 2 k∈I k
j

owing to the normalization condition (3). More generally, if Ij1 j2 …js is the set of
indices corresponding to the polynomials �k depending only on the parameters
�j1 , �j2 , … �js , the s-fold joint PCE-based Sobol’ indices are:

1 ∑
Sj1 j2 …js (�, t) = �k (uk (�, t))2 .
�2 k∈I j1 j2 …js

Conclusions

In this chapter, we have outlined the main ingredients of polynomial expansion meth-
ods for the pseudo-spectral analysis of random variables and fields, using either pro-
jections on orthonormal polynomials–the generalized polynomial chaos method, or
interpolations on Lagrange polynomials–the stochastic collocation method. We have
also shown how both approaches are actually intimately connected by a proper choice
of the integration/interpolation nodal sets used to compute the polynomial expansion
coefficients. However, alternative strategies have been recently considered in order to
evaluate them, which are detailed in the following chapters
“Generalized Polynomial Chaos for Non-intrusive Uncertainty Quantification in
Computational Fluid Dynamics” through “Screening Analysis and Adaptive Aparse
Collocation Methods”. Applications to uncertainty quantification and robust design
optimization for industrial challenges are given in parts III and IV.

References

1. Xiu, D., Karniadakis, G.: The Wiener-Askey polynomial chaos for stochastic differential equa-
tions. SIAM J. Sci. Comput. 24(2), 619–644 (2002)
2. Wiener, N.: The homogeneous chaos. Am. J. Math. 60(4), 897–936 (1938)
3. Xiu, D., Karniadakis, G.: Modeling uncertainty in flow simulations via generalized polynomial
chaos. J. Comput. Phys. 187(1), 137–167 (2003)
4. Lucor, D., Xiu, D., Su, C.H., Karniadakis, G.: Predictability and uncertainty in CFD. Int. J.
Numer. Methods Fluids 43(5), 483–505 (2003)
5. Le Maître, O., Knio, O., Najm, H., Ghanem, R.: A stochastic projection method for fluid flow.
I. Basic formulation. J. Comput. Phys. 173(2), 481–511 (2001)
6. Mathelin, L., Hussaini, M., Zang, T.: Stochastic approaches to uncertainty quantification in
CFD simulations. Numer. Algorithms 38(1), 209–236 (2005)
7. Walters, R., Huyse, L.: Uncertainty analysis for fluid mechanics with applications. ICASE
Report no. 2002-1 (2002)
8. Witteveen, J., Bijl, H.: Efficient quantification of the effect of uncertainties in advection-
diffusion problems using polynomial chaos. Numer Heat Transf Part B 53, 437–465 (2008)
9. Le Maître, O., Knio, O.M., Najm, H.N., Ghanem, R.G.: Uncertainty propagation using Wiener-
Haar expansions. J. Comput. Phys. 197(1), 28–57 (2004)
120 C. Lacor and Savin

10. Wan, X., Karniadakis, G.: An adaptive multi-element generalized polynomial chaos method
for stochastic differential equations. J. Comput. Phys. 209(2), 617–642 (2005)
11. Wan, X., Karniadakis, G.: Beyond Wiener-Askey expansions: handling arbitrary PDFs. J. Sci.
Comput. 27(1), 455–464 (2006)
12. Karhunen, K.: Zur spektraltheorie stochastischer prozesse. Ann. Acad. Sci. Fennicae 34, 1–7
(1946)
13. Loève, M.: Fonctions aleatoires du seconde ordre. In: Levy, P. (ed.) Processus Stochastiques
et Movement Brownien, Paris (1948)
14. Xiu, D.: Fast numerical methods for stochastic computations: a review. Commun. Comput.
Phys. 5(2–4), 242–272 (2009)
15. Dinescu, C., Smirnov, S., Hirsch, C., Lacor, C.: Assessment of intrusive and non-intrusive non-
deterministic CFD methodologies based on polynomial chaos expansions. Int. J. Eng. Syst.
Model. Simul. 2(1–2), 87–98 (2010)
16. Alekseev, A.K., Navon, I.M., Zelentsov, M.E.: The estimation of functional uncertainty using
polynomial chaos and adjoint equations. Int. J. Numer. Methods Fluids 67(3), 328–341 (2011)
17. Xiu, D., Tartakovsky, D.: Numerical methods for differential equations in random domains.
SIAM J. Sci. Comput. 28(3), 1167–1185 (2006)
18. Parussini, L., Pediroda, V.: Investigation of multi geometric uncertainties by different polyno-
mial chaos methodologies using a fictitious domain solver. Comput. Model. Eng. Sci. 23(1),
29–52 (2008)
19. Parussini, L.: Fictitious domain approach via Lagrange multipliers with least squares spectral
element method. J. Sci. Comput. 37(3), 316–335 (2008)
20. Perez, R., Walters, R.: An implicit polynomial chaos formulation for the Euler equations. In:
AIAA 2005-1406 (2005) 43rd AIAA Aerospace Sciences Meeting and Exhibit, 10–13 Jan
2005, Reno, NV
21. Perez, R.: Uncertainty analysis of computational fluid dynamics via polynomial chaos. Ph.D.
thesis, Virginia Polytechnic Institute and State University (2008)
22. Le Maître, O., Reagan, M., Najm, H., Ghanem, R., Knio, O.: A stochastic projection method
for fluid flow. II. Random process. J. Comput. Phys. 181(1), 9–44 (2002)
23. Reagan, M.T., Najm, H.N., Ghanem, R.G., Knio, O.M.: Uncertainty quantification in reacting-
flow simulations through non-intrusive spectral projection. Combust. Flame 132(3), 545–555
(2003)
24. Nobile, F., Tempone, R., Webster, C.G.: A sparse grid stochastic collocation method for partial
differential equations with random input data. SIAM J. Numer. Anal. 46(5), 2309–2345 (2008)
25. Berveiller, M., Sudret, B., Lemaire, M.: Stochastic finite element: a non-intrusive approach by
regression. Rev. Eur. Mécanique Numér. 15(1–3), 81–92 (2006)
26. Hosder, S., Walters, R.W., Perez, R.: A non-intrusive polynomial chaos method for uncertainty
propagation in CFD simulations. In: AIAA paper 2006-891 (2006) 44th AIAA Aerospace Sci-
ences Meeting and Exhibit, 9–12 Jan 2006, Reno, NV
27. Hosder, S., Walters, R.W., Balch, M.: Efficient sampling for non-intrusive polynomial chaos
applications with multiple uncertain input variables. In: AIAA paper 2007-1939 (2007) 48th
AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference,
23–26 April 2007, Honolulu, HI
28. Hosder, S., Walters, R.W., Balch, M.: Efficient uncertainty quantification applied to the aeroe-
lastic analysis of a transonic wing. In: AIAA paper 2008-729 (2008) 46th AIAA Aerospace
Sciences Meeting and Exhibit, 7–10 Jan 2008, Reno, NV
29. Smolyak, S.: Quadrature and interpolation formulas for tensor products of certain classes of
functions. Sov. Math. Dokl. 4, 240–243 (1963)
30. Gerstner, T., Griebel, M.: Numerical integration using sparse grids. Numer. Algorithms 18(3–
4), 209–232 (1998)
31. Novak, E., Ritter, K.: Simple cubature formulas with high polynomial exactness. Constr.
Approx. 15(4), 499–522 (1999)
32. Barthelmann, V., Novak, E., Ritter, K.: High dimensional polynomial interpolation on sparse
grids. Adv. Comput. Math. 12(4), 273–288 (2000)
General Introduction to Polynomial Chaos and Collocation Methods 121

33. Heiss, F., Winschel, V.: Likelihood approximation by numerical integration on sparse grids. J.
Econ. 144(1), 62–80 (2008)
34. Gerstner, T., Griebel, M.: Dimension-adaptive tensor-product quadrature. Computing 71(7),
65–87 (2003)
35. Ma, X., Zabaras, N.: An adaptive hierarchical sparse grid collocation algorithm for the solution
of stochastic differential equations. J. Comput. Phys. 228(8), 3084–3113 (2009)
36. Resmini, A., Peter, J., Lucor, D.: Sparse grids-based stochastic approximations with applica-
tions to aerodynamics sensitivity analysis. Int. J. Numer. Methods Eng. 106(1), 32–57 (2016)
37. Ghanem, R.: Hybrid stochastic finite elements and generalized monte carlo simulation. ASME
J. Appl. Mech. 65(4), 1004–1009 (1998)
38. Field, R.V.: Numerical methods to estimate the coefficients of the polynomial chaos expansion.
In: Proceedings of the 15th ASCE Engineering mechanics Conference (2002)
39. Ghiocel, D., Ghanem, R.: Stochastic finite-element analysis of seismic soil-structure interac-
tion. ASCE J. Eng. Mech. 128(1), 66–77 (2002)
40. Raisee, M., Kumar, D., Lacor, C.: A non-intrusive model reduction approach for polynomial
chaos expansion using proper orthogonal decomposition. Int. J. Numer. Methods Eng. 103(4),
293–312 (2015)
41. Tatang, M.A., Pan, W., Prinn, R.G., McRae, G.J.: An efficient method for parametric uncer-
tainty analysis of numerical geophysical models. J. Geophys. Res. 102(D18), 21925–21931
(1997)
42. Babuška, I., Nobile, F., Tempone, R.: A stochastic collocation method for elliptic partial dif-
ferential equations with random input data. SIAM J. Numer. Anal. 45(3), 1005–1034 (2007)
43. Ganapathysubramanian, B., Zabaras, N.: Sparse grid collocation schemes for stochastic natural
convection problems. J. Comput. Phys. 225(1), 652–685 (2007)
44. Xiu, D., Hesthaven, J.S.: High-order collocation methods for differential equations with ran-
dom inputs. SIAM J. Sci. Comput. 27(3), 1118–1139 (2005)
45. Xiu, D.: Numerical Methods for Stochastic Computations: A Spectral Method Approach.
Princeton University Press, Princeton, NJ (2010)
46. Eldred, M.S., Burkardt, J.: Comparison of non-intrusive polynomial chaos and stochastic col-
location methods for uncertainty quantification. In: AIAA paper 2009-0976 (2009) 47th AIAA
Aerospace Sciences Meeting, 5–8 Jan 2009, Orlando, FL
47. Loeven, G.J.A., Witteveen, J.A.S., Bijl, H.: Probabilistic collocation: An efficient non-intrusive
approach for arbitrarily distributed parametric uncertainties. In: AIAA paper 2007-317 (2007)
45th AIAA Aerospace Sciences Meeting and Exhibit, 8–11 Jan 2007, Reno, NV
48. Loeven, G.J.A., Bijl, H.: Airfoil analysis with uncertain geometry using the probabilistic collo-
cation method. In: AIAA paper 2008-2070 (2008) 49th AIAA/ASME/ASCE/AHS/ASC Struc-
tures, Structural Dynamics, and Materials Conference, 7–10 April 2008, Schaumburg, IL
49. Preston, J.S., Tasdizen, T., Terry, C.M., Cheung, A.K., Kirby, R.M.: Using the stochastic col-
location method for the uncertainty quantification of drug concentration due to depot shape
variability. IEEE Trans. Biomed. Eng. 56(3), 609–620 (2009)
50. Xiu, D.: Efficient collocational approach for parametric uncertainty analysis. Commun. Com-
put. Phys. 2(2), 293–309 (2007)
51. DeGennaro, A.M., Rowley, C.W., Martinelli, L.: Uncertainty quantification for airfoil icing
using polynomial chaos expansion. AIAA J. Aircr. 52(5), 1404–1411 (2015)
52. Wasilkowski, G.W., Woźniakowski, H.: Explicit cost bounds of algorithms for multivariate
tensor product problems. J. Complex. 11(1), 1–56 (1995)
53. Savin, É., Resmini, A., Peter, J.: Sparse polynomial surrogates for aerodynamic computa-
tions with random inputs. In: AIAA paper 2016-0433 (2016) 18th AIAA Non-Deterministic
Approaches Conference, 4–8 Jan 2016, San Diego, CA
54. Hesthaven, J.S., Gottlieb, S., Gottlieb, D.: Spectral Methods for Time-Dependent Problems.
Cambridge Monographs on Applied and Computational Mathematics, Vol. 21. Cambridge
University Press, Cambridge (2007)
55. Arnst, M., Ponthot, J.P.: An overview of nonintrusive characterization, propagation, and sen-
sitivity analysis of uncertainties in computational mechanics. Int. J. Uncertain. Quantif. 4(5),
387–421 (2014)
122 C. Lacor and Savin

56. Eldred, M.S.: Design under uncertainty employing stochastic expansion methods. Int. J. Uncer-
tain. Quantif. 1(2), 119–146 (2011)
57. Sudret, B.: Global sensitivity analysis using polynomial chaos expansions. Reliab. Eng. Syst.
Saf. 93(7), 964–979 (2008)