Abstract
Digital twin models allow us to continuously assess the possible risk of damage and failure of a complex system. Yet highfidelity digital twin models can be computationally expensive, making quickturnaround assessment challenging. Toward this goal, this article proposes a novel bifidelity method for estimating the conditional valueatrisk (CVaR) for nonlinear systems subject to dependent and highdimensional inputs. For models that can be evaluated fast, a method that integrates the dimensionally decomposed generalized polynomial chaos expansion (DDGPCE) approximation with a standard samplingbased CVaR estimation is proposed. For expensivetoevaluate models, a new bifidelity method is proposed that couples the DDGPCE with a Fourierpolynomial expansion of the mapping between the stochastic lowfidelity and highfidelity output data to ensure computational efficiency. The method employs measureconsistent orthonormal polynomials in the random variable of the lowfidelity output to approximate the highfidelity output. Numerical results for a structural mechanics truss with 36dimensional (dependent random variable) inputs indicate that the DDGPCE method provides very accurate CVaR estimates that require much lower computational effort than standard GPCE approximations. A second example considers the realistic problem of estimating the risk of damage to a fiberreinforced composite laminate. The highfidelity model is a finite element simulation that is prohibitively expensive for risk analysis, such as CVaR computation. Here, the novel bifidelity method can accurately estimate CVaR as it includes lowfidelity models in the estimation procedure and uses only a few highfidelity model evaluations to significantly increase accuracy.
This is a preview of subscription content, access via your institution.
Notes
The GPCE in (6) should not be confused with that of (Xiu and Karniadakis 2002). The GPCE, presented here, is meant for an arbitrary dependent probability distribution of random input. In contrast, the existing PCE, whether classical (Wiener 1938) or generalized (Xiu and Karniadakis 2002), still requires independent random inputs.
References
Artzner P, Delbaen F, Eber JM, Heath D (1999) Coherent measures of risk. Math Finance 9(3):203–228
Bernal F, Gobet E, Printems J (2020) Volatility uncertainty quantification in a stochastic control problem applied to energy. Methodol Comput Appl Probab 22(1):135–159
Brunton SL, Nathan Kutz J, Manohar K, Aravkin AY, Morgansen K, Klemisch J, Goebel N, Buttrick J, Poskin J, BlomSchieber AW, Hogan T, McDonard D (2021) Datadriven aerospace engineering: reframing the industry with machine learning. AIAA J 59(8):2820–2847
Chaudhuri A, Peherstorfer B, Willcox K (2020) Multifidelity crossentropy estimation of conditional valueatrisk for riskaverse design optimization. In: AIAA Scitech 2020 Forum, p 2129
Chaudhuri A, Kramer B, Norton M, Royset JO, Willcox K (2022) Certifiable riskbased engineering design optimization. AIAA J 60(2):551–565
Chen X, Nelson BL, Kim KK (2012) Stochastic kriging for conditional valueatrisk and its sensitivities. In: Proceedings of the 2012 Winter Simulation Conference (WSC), IEEE, pp 1–12
Duarte A, Díaz Sáez A, Silvestre N (2017) Comparative study between XFEM and Hashin damage criterion applied to failure of composites. ThinWalled Struct 115:277–288
Garreis S, Surowiec TM, Ulbrich M (2021) An interiorpoint approach for solving riskaverse PDEconstrained optimization problems with coherent risk measures. SIAM J Optim 31(1):1–29
Jy G, Uryasev S (2017) Support vector machines based on convex risk functions and general norms. Ann Oper Res 249(1):301–328
Hadigol M, Doostan A (2018) Least squares polynomial chaos expansion: a review of sampling strategies. Comput Methods Appl Mech Eng 332:382–407
Heinkenschloss M, Kramer B, Takhtaganov T, Willcox K (2018) Conditionalvalueatrisk estimation via reducedorder models. SIAM/ASA J Uncertain Quantif 6(4):1395–1423
Heinkenschloss M, Kramer B, Takhtaganov T (2020) Adaptive reducedorder model construction for conditional valueatrisk estimation. SIAM/ASA J Uncertain Quantif 8(2):668–692
Hong LJ, Hu Z, Liu G (2014) Monte Carlo methods for valueatrisk and conditional valueatrisk: a review. ACM Trans Model Comput Simul (TOMACS) 24(4):1–37
Jakeman JD, Franzelin F, Narayan A, Eldred M, Pflüger D (2019) Polynomial chaos expansions for dependent random variables. Comput Methods Appl Mech Eng 351:643–666
Jakeman JD, Kouri DP, Huerta JG (2022) Surrogate modeling for efficiently, accurately and conservatively estimating measures of risk. Reliab Eng Syst Saf 221:108280
Kennedy MC, O’Hagan A (2000) Predicting the output from a complex computer code when fast approximations are available. Biometrika 87(1):1–13
Kouri DP, Surowiec TM (2016) Riskaverse PDEconstrained optimization using the conditional valueatrisk. SIAM J Optim 26(1):365–396
Kouri DP, Surowiec TM (2020) Epiregularization of risk measures. Math Oper Res 45(2):774–795
Lattanzi L, Raffaeli R, Peruzzini M, Pellicciari M (2021) Digital twin for smart manufacturing: a review of concepts towards a practical industrial implementation. Int J Comput Integr Manuf 34(6):567–597
Lee D (2021) Stochastic optimization for design under uncertainty with dependent random variables. PhD thesis, The University of Iowa
Lee D, Rahman S (2020) Practical uncertainty quantification analysis involving statistically dependent random variables. Appl Math Model 84:324–356
Lee D, Rahman S (2021) Robust design optimization under dependent random variables by a generalized polynomial chaos expansion. Struct Multidisc Optim 63(5):2425–2457
Lee D, Rahman S (2022) Reliabilitybased design optimization under dependent random variables by a generalized polynomial chaos expansion. Struct Multidisc Optim 65(1):1–29
Luthen N, Marelli S, Sudret B (2021) Sparse polynomial chaos expansions: literature survey and benchmark. SIAM/ASA J Uncertain Quantif 9(2):593–649
Mansini R, Ogryczak W, Speranza MG (2007) Conditional value at risk and related linear programming models for portfolio optimization. Ann Oper Res 152:227–256
MATLAB (2021) version 9.10.0.1613233 (R2021b). The MathWorks Inc., Natick, Massachusetts
Moure M, SanchezSaez S, Barbero E, Barbero E (2014) Analysis of damage localization in composite laminates using a discrete damage model. Composites B 66:224–232
Navarro M, Witteveen JAS, Blom JG (2014) Polynomial chaos expansion for general multivariate distributions with correlated variables. arXiv:1406.5483
Niederer SA, Sacks MS, Girolami M, Willcox K (2021) Scaling digital twins from the artisanal to the industrial. Nat Comput Sci 1(5):313–320
Noh Y, Choi KK, Du L (2009) Reliabilitybased design optimization of problems with correlated input variables using a Gaussian copula. Struct Multidisc Optim 38:1–16
Norton M, Khokhlov V, Uryasev S (2021) Calculating cvar and bpoe for common probability distributions with application to portfolio optimization and density estimation. Ann Oper Res 299(1):1281–1315
Peherstorfer B, Willcox K, Gunzburger M (2016) Optimal model management for multifidelity Monte Carlo estimation. SIAM J Sci Comput 38(5):A3163–A3194
Peherstorfer B, Willcox K, Gunzburger M (2018) Survey of multifidelity methods in uncertainty propagation, inference, and optimization. SIAM Rev 60(3):550–591
Pepper N, Gaymann A, Sharma S, Montomoli F (2021) Local bifidelity field approximation with knowledge based neural networks for computational fluid dynamics. Sci Rep 11(1):1–11
Rabitz H, Aliş Ö, Shorter J, Shim K (1999) Efficient input–output model representations. Comput Phys Commun 117(1):11–20
Rahman S (2008) A polynomial dimensional decomposition for stochastic computing. Int J Numer Meth Eng 76(13):2091–2116
Rahman S (2018) A polynomial chaos expansion in dependent random variables. J Math Anal Appl 464(1):749–775
Rahman S (2019) Uncertainty quantification under dependent random variables by a generalized polynomial dimensional decomposition. Comput Methods Appl Mech Eng 344:910–937
Rockafellar RT, Royset JO (2015) Engineering decisions under risk averseness. ASCEASME J Risk Uncertain Eng Syst Part A 1(2):04015003
Rockafellar RT, Uryasev S (2002) Conditional valueatrisk for general loss distributions. J Bank Finance 26(7):1443–1471
Rockafellar RT, Uryasev S et al (2000) Optimization of conditional valueatrisk. J Risk 2:21–42
Royset JO, Bonfiglio L, Vernengo G, Brizzolara S (2017) Riskadaptive setbased design and applications to shaping a hydrofoil. J Mech Des 139(10):101403
Sarykalin S, Serraino G, Uryasev S (2008) Valueatrisk vs. conditional valueatrisk in risk management and optimization. In: Stateoftheart decisionmaking tools in the informationintensive age, INFORMS, pp 270–294
Sobol IM (1967) Distribution of points in a cube and approximate evaluation of integrals. Zh Vych Mat Mat Fiz 7:784–802
Soma T, Yoshida Y (2020) Statistical learning with conditional value at risk. arXiv:2002.05826
Tan SC (1991) A progressive failure model for composite laminates containing openings. J Compos Mater 25(5):556–577
Wiener N (1938) The homogeneous chaos. Am J Math 60(4):897–936
Xiu D, Karniadakis GE (2002) The Wiener–Askey polynomial chaos for stochastic differential equations. SIAM J Sci Comput 24:619–644
Yang H, Gunzburger M (2017) Algorithms and analyses for stochastic optimization for turbofan noise reduction using parallel reducedorder modeling. Comput Methods Appl Mech Eng 319:217–239
Zou Z, Kouri DP, Aquino W (2019) An adaptive local reduced basis method for solving PDEs with uncertain inputs and evaluating risk. Comput Methods Appl Mech Eng 345:302–322
Funding
This material is based on research sponsored by the Air Force Research Lab (AFRL) and Defense Advanced Research Projects Agency (DARPA) under agreement number FA86502127126. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Replication of results
The MATLAB code files of the Examples 1–2 are provided in the following Github repository: https://github.com/icdsigmalee/BifidelityCVaRestimationbyDDGPCE.git. Unfortunately, the source codes implementing a set of builtin Abaqus functions for FEA, used in Example 2, cannot be shared due to the external code license. Readers further interested in the codes are encouraged to contact the authors by email.
Disclaimer
The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the Air Force Research Lab (AFRL) and Defense Advanced Research Projects Agency (DARPA) or the U.S. Government.
Additional information
Responsible Editor: Chao Hu
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Topical Collection: Advanced Optimization Enabling Digital Twin Technology.
Guest Editors: C Hu, VA González, T Kim, O San, Z Hu, P Zheng
Appendix: Three step process to construct measureconsistent orthonormal polynomials
Appendix: Three step process to construct measureconsistent orthonormal polynomials
This appendix summarizes a process to generate the multivariate orthonormal polynomial basis of GPCE in Sect. 2.3. The orthonormal polynomial functions are consistent with an arbitrary, nonproducttype probability measure \(f_{\textbf{X}}(\textbf{x})\text {d}\textbf{x}\) of \(\textbf{X}\) and determined by the following three steps.

1.
Given \(m \in \mathbb {N}_0\), create an \(L_{N,m}\)dimensional column vector
$$\begin{aligned} \textbf{M}_m(\textbf{x})=(\textbf{x}^{\textbf{j}^{(1)}},\ldots ,\textbf{x}^{\textbf{j}^{(L_{N,m})}})^{\intercal }, \end{aligned}$$of monomials whose elements are the monomials \(\textbf{x}^{\textbf{j}}\) for \(\textbf{j}\le m\) arranged in the aforementioned order. It is referred to as the monomial vector in \(\textbf{x}=(x_1,\ldots ,x_N)^\intercal\) of degree at most m.

2.
Construct an \(L_{N,m} \times L_{N,m}\) monomial moment matrix of \(\textbf{M}_m(\textbf{X})\), defined as
$$ {\textbf{G}}_m:= \mathop {\mathrm {\mathbb {E}}}\limits [\textbf{M}_m(\textbf{X})\textbf{M}_{m}^{\intercal }(\textbf{X})]=\int _{\mathbb {{\bar{A}}}^{N}}\textbf{M}_{m}(\textbf{x})\textbf{M}_{m}^{\intercal }(\textbf{x}) f_{\textbf{X}}(\textbf{x})\text {d}.$$For an arbitrary PDF \(f_{\textbf{X}}(\textbf{x})\), \(\textbf{G}_m\) cannot be determined exactly, but it can be estimated with good accuracy by numerical integration and/or sampling methods (Lee and Rahman 2020).

3.
Select the \(L_{N,m} \times L_{N,m}\) whitening matrix \({\textbf{W}}_m\) from the Cholesky decomposition of the monomial moment matrix \({\textbf{G}}_m\) (Rahman 2018), leading to
$$\begin{aligned} {\textbf{W}}_{m}^{1}{\textbf{W}}_{m}^{\intercal }={\textbf{G}}_{m}. \end{aligned}$$Then employ the whitening transformation to generate multivariate orthonormal polynomials from
$$\begin{aligned} {\varvec{\Psi }}_m(\textbf{x})={\textbf{W}}_m \textbf{M}_m(\textbf{x}). \end{aligned}$$
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author selfarchiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Lee, D., Kramer, B. Bifidelity conditional valueatrisk estimation by dimensionally decomposed generalized polynomial chaos expansion. Struct Multidisc Optim 66, 33 (2023). https://doi.org/10.1007/s00158022034776
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00158022034776
Keywords
 Risk measures
 Conditional valueatrisk
 Generalized polynomial chaos expansion
 Dimensionally decomposed GPCE
 Bifidelity modeling