Pith. sign in

REVIEW 3 major objections 5 minor 60 references

Optimal Uncertainty Quantification under General Moment Constraints on Input Subdomains

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper proves that failure-probability bounds under subdomain moment constraints reduce to an unconstrained optimization over free canonical moments, with evidence theory emerging when only subdomain masses are known.

desk verdict Subdomain-moment OUQ is a genuinely useful idea, but the normalization error in Eq. (17) breaks the central reformulation and the evidence-theory equivalence rests on a false product-form assumption. read the letter →

arxiv 2512.19572 v2 pith:5GA5PAPM submitted 2025-12-22 physics.comp-ph

Rong Jin , Xingsheng Sun This is my paper · ORCID
classification physics.comp-ph
keywords optimaluncertaintyquantificationtruncatedmomentssubdomainmomentconstraintscanonicalevidencetheoryprobabilityoffailureinversetransformsamplingsafetycertification
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that optimal bounds on the probability of failure can be computed rigorously when the only available input information is a set of truncated moment constraints defined on subdomains of each input variable. It proves that the infinite-dimensional optimization over admissible probability measures collapses to a finite-dimensional unconstrained search over free canonical moments, with the Dirac supports and weights recovered from an eigendecomposition that exactly satisfies the moment constraints. It further shows that when only zeroth-order moments (subdomain masses) are specified, the resulting bounds coincide with plausibility and belief from evidence theory. If correct, this gives safety certification a way to use partial, local data—conditional means, variances, higher moments—without committing to a full distribution, and the numerical results show the bounds tighten systematically as subdomains or moment orders increase.

What carries the argument

Free canonical moments. On each subdomain, the prescribed moments are transformed to canonical moments in [0,1]; adding r+1 free higher-order canonical moments produces a full sequence from which a Jacobi matrix is built. The matrix's eigendecomposition yields the Dirac supports and weights, ensuring exact moment matching and admissible nonnegative weights. Inverse transform sampling (ITS) optionally estimates the PoF by drawing from the resulting discrete CDF, which the paper shows costs up to two orders of magnitude less than exhaustive enumeration while keeping relative error under 1%.

What would settle it

Compute the true worst-case failure probability over product measures with fixed subdomain masses for a two-input sum model G(X1,X2)=X1+X2 with output threshold c, and compare it to the closed-form upper bound from Eq. (33). If the closed-form bound is lower than the true supremum, the product-failure assumption fails; the same comparison can be run in low dimension with the paper's own optimization machinery.

Watch

Extended reading notes

Core claim

The central claim is that the subdomain-moment OUQ problem is exactly solvable by reduction. For each input and subdomain, the r+1 constrained raw moments fix the first r+1 canonical moments; the remaining r+1 free canonical moments, varying in (0,1), parameterize all admissible measures. The extremal measures are discrete (Dirac) with r+1 points per subdomain, and the paper constructs them via an eigendecomposition of a symmetric tridiagonal Jacobi matrix, whose eigenvalues are the supports and whose normalized first eigenvector components give the weights. The paper also claims that in the zeroth-order case the bounds are given by closed-form sums over subdomains: the upper bound equals pl

Load-bearing premise

The equivalence to evidence theory, and the closed-form bounds of Eq. (33), rests on the assumption that the failure set in input space is a Cartesian product of per-input failure sets because the inputs are independent—false for typical coupled forward models like G(X1,X2)=X1+X2.

Editorial extensions

If this is right

  • If correct, an engineer can certify safety on the basis of local moment data (subdomain masses, conditional means, variances) without assuming a PDF; the verdict is rigorous: safe if the upper bound is at or below the tolerance, unsafe if the lower bound is above it.
  • The numerical results establish that adding subdomains or higher-order moment constraints monotonically narrows the [L,U] interval, and in the 5D nonlinear and ballistic examples the interval can be driven to within a few percent of the true PoF with 8 subdomains and second-order moments.
  • The equivalence to evidence theory means belief and plausibility are the optimal OUQ bounds under subdomain mass constraints, linking two previously separate uncertainty frameworks.
  • In high dimensions, ITS plus common random numbers keeps the search tractable, with reported speedups of up to two orders of magnitude and errors below 1% at 50,000 samples.
  • The ballistic example shows parameter-specific sensitivity: tightening bounds for the yield stress A is better done by refining subdomain partitioning, while for hardening modulus B higher-order moments matter differently—guiding where to spend data-collection effort.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The product-failure assumption in the zeroth-order derivation (Eq. 32) is likely false for many realistic forward maps; testing the closed-form bounds against direct optimization on a non-product failure set (e.g., a sum threshold) would reveal whether the evidence-theory equivalence is general or only holds under a separability condition.
  • The canonical-moment parameterization could enable adaptive subdomain strategies, since the paper notes that subdomains intersecting the failure boundary are the ones that actually tighten bounds.
  • A natural extension is mixed aleatory/epistemic settings where some inputs have full stochastic descriptions and others are known only by subdomain moments; the reduction machinery should carry over, and a rigorous comparison of the resulting bounds against Monte Carlo would be a testable next step.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes an optimal uncertainty quantification (OUQ) framework for bounding the probability of failure when input distributions are known only through truncated raw moments on subdomains of the input space. The authors invoke the reduction theorem to represent admissible measures as Dirac mixtures, then introduce a canonical-moment parameterization that converts the constrained infinite-dimensional problem into an unconstrained finite-dimensional optimization over free canonical moments. They further claim that in the zeroth-order moment case the OUQ formulation is equivalent to evidence theory, with bounds given by plausibility and belief. The methodology is illustrated on one-dimensional identity functions, a five-dimensional nonlinear function, and a ballistic impact problem with a neural-network surrogate.

Significance. The idea of adapting canonical moments to subdomain-local moment constraints is attractive and, if implemented correctly, could yield a practical alternative to nested-loop or penalty-based OUQ solvers. The paper is clearly written, the numerical studies are extensive, and the surrogate modeling for the ballistic problem is carefully validated. However, the central mathematical construction contains a normalization error that invalidates the parameterization for K>1, and the evidence-theory equivalence rests on a false product-structure assumption for multi-dimensional failure domains. As written, the principal claims of rigorous optimal bounds and of an evidence-theory equivalence are not established. The manuscript could become a useful contribution after a substantial revision that corrects these issues and re-runs the numerical experiments.

major comments (3)
  1. [§2.4.1–2.4.4, Eq. (17) and Eq. (24)] The affine transformation of moments does not normalize the subdomain mass. For k=1, Eq. (17) gives M̃_{i,j,0}=M_{i,j,0}, which equals 1 only when the subdomain carries the entire probability. Canonical moments are defined for probability measures with total mass 1, so the Jacobi-matrix construction (Eqs. (19)–(23)) produces a unit-mass measure whose moments are the unnormalized M̃, not the conditional moments M̃/M_{i,j,0}. Multiplying by M_{i,j,0} in Eq. (24) then yields a measure with moments M_{i,j,0} times the unit-mass moments, which do not match the prescribed M_{i,j,k-1} for k>1. For example, for a uniform measure on [0,1] restricted to [0,0.5], M=(0.5,0.125,1/24); Eq. (17) gives M̃=(0.5,0.25,1/6), and the recovery step returns subdomain moments (0.5,0.125,1/12) rather than the prescribed (0.5,0.125,1/24). Thus the objective (25) optimizes over a set different from the admissible
  2. [§3 and Appendix A, Eq. (32)] The assumption that Ω^c = ∏ Ω_i^c is not a consequence of independence of the inputs. Independence concerns the product structure of the probability measure, not the geometry of the failure region. For a general forward map, e.g., G(X1,X2)=X1+X2 with failure {G≥c}, Ω^c is a half-plane, which is not a Cartesian product. Therefore Eq. (33) does not give the OUQ upper and lower bounds for general multi-dimensional problems, and the claimed equivalence to evidence theory does not hold as stated. The derivation in Appendix A relies entirely on the product factorization. The equivalence may be valid for one-dimensional inputs or for product-form failure domains, but the paper applies it to the five-dimensional and ballistic-impact examples, where it is unjustified.
  3. [§2.4.6 and §4.2–4.3] In cases where (K(r+1))^m ≥ N_ITS, the PoF is replaced by a Monte Carlo estimate via Eq. (28). The resulting objective is noisy, and the differential-evolution search is run on this stochastic objective. The abstract and Section 2.1 claim 'rigorous optimal bounds,' but for the high-dimensional examples the reported bounds are statistical estimates, not rigorous bounds. The paper should qualify this claim, for instance by providing conservative confidence intervals or by clearly separating rigorous low-dimensional results from approximate high-dimensional ones.
minor comments (5)
  1. [Eq. (13)] The notation {x} ∈ ℝ^{m k (r+1)} uses a lowercase k, which is inconsistent with the number of subdomains K; it should be mK(r+1).
  2. [§2.2] The statement that all inputs share the same number of available moments r is slightly confusing because the constraints are listed for k=1,...,r+1 (moments of order 0,...,r). Clarify the terminology.
  3. [Eq. (18)] After correcting the normalization, the formulas for canonical moments should be expressed in terms of the normalized moments (i.e., conditional moments divided by M_{i,j,0}) to avoid ambiguity.
  4. [Appendix B, Eq. (51)] In Eq. (51), the subscript of the supremum should be μ∈A_j rather than μ∈A; the proof defines A_j but then uses A in the display.
  5. [Figure 3] For the lower bound in the figure, it would be clearer to explicitly indicate which subdomains are fully contained in the failure region rather than relying on the reader's interpretation of the shaded areas.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central reformulation and bounds are derived from stated moment constraints and external reduction/canonical-moment theory, not from the target PoF.

full rationale

The paper's central claim is that subdomain truncated-moment constraints define an admissible set (Eq. 9) and that OUQ bounds over that set can be computed via reduction to Dirac measures (Eq. 13), canonical-moment reparameterization (Eqs. 17-25), and eigendecomposition (Eqs. 19-23). The reduction theorem is cited to Ref. [17] (Owhadi et al.), and the canonical-moment machinery to Refs. [34] and [20]; none of these are the present authors' work, and the constraints are not defined in terms of the PoF. The free canonical moments are optimization variables representing higher-order moments of admissible measures; they are not fitted to the PoF and then reported as a prediction. In the numerical examples, moments are computed from the true distributions only as benchmark inputs and the true PoF is computed independently, so the bracketing of the true PoF is a self-contained validation. The evidence-theory equivalence (Section 3) is derived, not assumed: the paper starts from zeroth-moment constraints and derives bounds (Eq. 33) whose form matches plausibility/belief. Whether the factorization of the failure domain in Eq. (32) is valid for general G is a correctness/assumption concern, not a circular one; the same is true for the affine-moment normalization in Eq. (17), which affects whether the canonical-moment construction enforces the stated moments but is not an input-output identity. Self-citations (e.g., Refs. [22], [37], [60]) are used for background or motivation and are not load-bearing for the central derivation. I therefore find no step where a prediction is equivalent by construction to its inputs.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The method introduces no new physical entities. It relies on established mathematical theorems (reduction, canonical moments, quadrature), the domain assumption that the specified truncated moments bound the true measure, and the problematic product-form failure-domain assumption that drives the evidence-theory equivalence.

free parameters (2)
  • N_ITS (inverse transform sample count) = 5 × 10^4
    Chosen from convergence tests in Section 4.2; determines the claimed <1% relative error and the two-order-of-magnitude speedup.
  • DE population size and max iterations = 20–50 × n_vars; 100–200 iterations
    Heuristic settings stated in Section 2.5; no guarantee of global optimality; the reported bounds depend on them.
assumptions (6)
  • standard math Reduction theorem: extremal measures for subdomain moment constraints are discrete with (r+1) points per subdomain per input (Eqs. 11–13).
    Invoked from Owhadi et al. [17]; the extension from global to per-subdomain constraints is asserted, not proved in this paper.
  • standard math Canonical moment representation and Golub-Welsch quadrature: any admissible canonical-moment sequence yields a positive Dirac measure matching the fixed moments (Sections 2.4.2–2.4.4).
    Standard canonical moment theory [34,38]; assumed without derivation.
  • domain assumption The true measure lies in the admissible set defined by the truncated moments (Eq. 9).
    Needed for the certification logic in Section 2.1; cannot be verified from data.
  • domain assumption Inputs are independent: μ = ⊗ μ_i (Eq. 9).
    Used throughout; independence justifies the product-form Dirac measures and per-marginal inverse transform sampling.
  • ad hoc to paper Failure domain decomposes as X^c = ∏ X_i^c (Eq. 32).
    Assumed in Section 3 and Appendix A; false for a general forward map G such as G = X1 + X2, so the evidence-theory equivalence as stated does not hold for general G.
  • domain assumption Neural-network surrogate S_θ accurately reproduces the LS-DYNA output (Section 4.3.2).
    Ballistic-impact 'true PoF' and bounds are computed through the surrogate, not the original simulator; R^2 ≈ 1.00 on the test set, but the surrogate itself is not certified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Uncertainty Quantification under General Moment Constraints on Input Subdomains." pith.science (2026). https://pith.science/paper/5GA5PAPM

@misc{pith2026251219572,
  author       = {Pith},
  title        = {Pith review of: Optimal Uncertainty Quantification under General Moment Constraints on Input Subdomains},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5GA5PAPM}},
  note         = {Machine review of arXiv:2512.19572}
}
read the original abstract

We present an optimal uncertainty quantification (OUQ) framework for systems whose uncertain inputs are characterized by truncated moment constraints defined over subdomains. Based on this partial information, rigorous optimal upper and lower bounds on the probability of failure (PoF) are derived over the admissible set of probability measures, providing a principled basis for system safety certification. We formulate the OUQ problem under general subdomain moment constraints and develop a high-performance computational framework to compute the optimal bounds. This approach transforms the original infinite-dimensional optimization problems into finite-dimensional unconstrained ones parameterized solely by free canonical moments. To address the prohibitive cost of PoF evaluation in high-dimensional settings, we incorporate inverse transform sampling (ITS), enabling efficient and accurate PoF estimation within the OUQ optimization. We also demonstrate that constraining inputs only by zeroth-order moments over subdomains yields a formulation equivalent to evidence theory. Three groups of numerical examples demonstrate the framework's effectiveness and scalability. Results show that increasing the number of subdomains or the moment order systematically tightens the bound interval. For high-dimensional problems, the ITS strategy reduces computational costs by up to two orders of magnitude while maintaining relative error below 1%. Furthermore, we identify regimes where optimal bounds are sensitive to subdomain partitioning or higher-order moments, guiding uncertainty reduction efforts for safety certification.

Figures

Figures reproduced from arXiv: 2512.19572 by the authors.

Figure 1
Figure 1. Illustration of certification in OUQ: (a) Rigorous decision criterion for system safety, and (b) Convergence of optimal bounds with increasing information. 2.2. OUQ under Local Moment Constraints In this work, we consider a scenario in which the input domain of the system, , is partitioned into multiple subdomains. Without loss of generality, we assume that the domain of the 𝑖-th input variable, 𝑖 , is divided int… view at source ↗
Figure 2
Figure 2. Illustration of ITS in OUQ: (a) PDF of a three-point Dirac measure, and (b) Corresponding CDF and ITS. An additional motivation for using ITS arises from examining how each term of the joint Dirac measure contributes to the PoF. As shown in Eq. (25), the Dirac supports determine whether a term contributes, while the Dirac weights determine the magnitude of that contribution. The contribution of a joint term is on th… view at source ↗
Figure 3
Figure 3. Schematic illustration of the probability constraints for a one-dimensional random variable. The input domain is partitioned into subintervals 1,𝑗, and the failure region  c 1 is highlighted for comparison. In this example, the upper and lower bounds on the PoF are 𝑈 = 𝑀1,2,0 + 𝑀1,3,0 + 𝑀1,4,0 + 𝑀1,5,0 and 𝐿 = 𝑀1,3,0 + 𝑀1,4,0 , respectively. Since the random inputs are independent, the failure domain can be decomp… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Truncated PDFs and corresponding failure regions for the four considered distributions: (a) truncated normal, (b) uniform, (c) truncated Weibull, and (d) truncated bimodal normal-mixture. A clear trend observed in [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Optimal PoF bounds as functions of the number of subdomains and moment constraints: (a) truncated normal, (b) uniform, (c) truncated Wei-bull, and (d) truncated bimodal normal-mixture. An additional observation in [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: Comparison between the Dirac CDFs at the optimal bounds and the corresponding true CDFs: (a) truncated normal, (b) uniform, (c) truncated Wei-bull, and (d) truncated bimodal normal mixture. 4.2. Case 2: A Five-Dimensional Nonlinear Problem We next consider a more compl…
Figure 7
Figure 7. Figure 7: Estimation of PoF using ITS: (a) 𝐾 = 4, 𝑟 = 2; and (b) 𝐾 = 8, 𝑟 = 2. via ITS converges to the exact value as 𝑁ITS increases. In particular, when 𝑁ITS > 4 × 104 , the approximation error falls below 1%. For the second case, [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: DE convergence under different situations: (a) 𝐾 = 1, 𝑟 = 0; (b) 𝐾 = 1, 𝑟 = 1; (c) 𝐾 = 1, 𝑟 = 2; (d) 𝐾 = 2, 𝑟 = 0; (e) 𝐾 = 2, 𝑟 = 1; (f) 𝐾 = 2, 𝑟 = 2; (g) 𝐾 = 4, 𝑟 = 0; (h) 𝐾 = 4, 𝑟 = 1; (i) 𝐾 = 4, 𝑟 = 2; (j) 𝐾 = 8, 𝑟 = 0; (k) 𝐾 = 8, 𝑟 = 1; and (l) 𝐾 = 8, 𝑟 = 2. R. Jin…
Figure 9
Figure 9. Figure 9: Optimal PoF bounds for the five-dimensional nonlinear problem. Another notable observation from [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: Schematic illustration of an AZ31B magnesium plate impacted by a spherical steel projectile at 200 m/s: (a) Initial configuration, and (b) Performance metric, with the maximum backface deflection denoted by 𝑌 . In each subfigure, the top figure shows a three-dimension…
Figure 11
Figure 11. Figure 11: Surrogate model performance: (a) Training and validation MSE versus epoch on a logarithmic scale, and (b) Parity plot comparing predicted and true outputs on the test set. To quantify the computational acceleration provided by the surrogate forward model, we benchmark…
Figure 12
Figure 12. Figure 12: Optimal PoF bounds for the ballistic impact problem. A previous study has shown that, when only the admissible ranges of the JC parameters (𝐴, 𝐵, 𝑛, 𝐶, 𝑚) are specified, the maximum backface deflection 𝑌 is most sensitive to the quasi-static yield stress 𝐴 [60]. That …
Figure 13
Figure 13. Figure 13: Optimal PoF upper bounds for one-dimensional OUQ analyses of the JC parameters: (a) quasi-static yield stress 𝐴, and (b) hardening modulus 𝐵, for three cases with identical computational cost. 5. Concluding Remarks In this study, we have formulated an OUQ framework to…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 3 linked inside Pith

  1. [1]

    Kovachki, B

    N. Kovachki, B. Liu, X. Sun, H. Zhou, K. Bhattacharya, M. Ortiz, A. Stuart, Multiscale modeling of materials: Computing, data science, uncertainty and goal-oriented optimization, Mechanics of Materials 165 (2022) 104156

    2022

  2. [2]

    B.Liu,X.Sun,K.Bhattacharya,M.Ortiz, Hierarchicalmultiscalequantificationofmaterialuncertainty, JournaloftheMechanicsandPhysics of Solids 153 (2021) 104492

    2021

  3. [3]

    R. Jin, G. Wang, X. Sun, Ensemble-based data assimilation for material model characterization in high-velocity impact, arXiv preprint arXiv:2510.09703 (2025)

    arXiv 2025

  4. [4]

    Der Kiureghian, O

    A. Der Kiureghian, O. Ditlevsen, Aleatory or epistemic? does it matter?, Structural safety 31 (2009) 105–112

    2009

  5. [5]

    W. K. Hastings, Monte carlo sampling methods using markov chains and their applications (1970)

    1970

  6. [6]

    A.P.Dempster, Ageneralizationofbayesianinference, JournaloftheRoyalStatisticalSociety:SeriesB(Methodological)30(1968)205–232

    1968

  7. [7]

    T. Fetz, M. Oberguggenberger, Propagation of uncertainty through multivariate functions in the framework of sets of probability measures, Reliability Engineering & System Safety 85 (2004) 73–87

    2004

  8. [8]

    Zhang, M

    J. Zhang, M. D. Shields, The effect of prior probabilities on quantification and propagation of imprecise probabilities resulting from small datasets, Computer Methods in Applied Mechanics and Engineering 334 (2018) 483–506. R. Jin et al.:Preprint submitted to ElsevierPage 27 of 31 Table 6 Optimal bounds and relative deviations for the one-dimensional tru...

    2018

Show all 60 references
  1. [9]

    Zhang, M

    J. Zhang, M. D. Shields, On the quantification and efficient propagation of imprecise probabilities resulting from small datasets, Mechanical Systems and Signal Processing 98 (2018) 465–483

    2018

  2. [10]

    A. P. Dempster, Upper and lower probabilities induced by a multivalued mapping, in: Classic works of the Dempster-Shafer theory of belief functions, Springer, 2008, pp. 57–72

    2008

  3. [11]

    Shafer, A mathematical theory of evidence (2020)

    G. Shafer, A mathematical theory of evidence (2020)

    2020

  4. [12]

    Ferson, J

    S. Ferson, J. G. Hajagos, Arithmetic with uncertain numbers: rigorous and (often) best possible answers, Reliability Engineering & System Safety 85 (2004) 135–152

    2004

  5. [13]

    Weichselberger, The theory of interval-probability as a unifying concept for uncertainty, International Journal of Approximate Reasoning 24 (2000) 149–170

    K. Weichselberger, The theory of interval-probability as a unifying concept for uncertainty, International Journal of Approximate Reasoning 24 (2000) 149–170. R. Jin et al.:Preprint submitted to ElsevierPage 28 of 31 Table 8 Optimal bounds and relative deviations for the one-d...

    2000

  6. [14]

    J. J. Buckley, Fuzzy probabilities: new approach and applications, Springer, 2005

    2005

  7. [15]

    M. Beer, S. Ferson, V. Kreinovich, Imprecise probabilities in engineering analyses, Mechanical systems and signal processing 37 (2013) 4–29

    2013

  8. [16]

    M.Götz,W.Graf,M.Kaliske, Structuraldesignwithpolymorphicuncertaintymodels, InternationalJournalofReliabilityandSafety9(2015) 112–131

    2015

  9. [17]

    Owhadi, C

    H. Owhadi, C. Scovel, T. J. Sullivan, M. McKerns, M. Ortiz, Optimal uncertainty quantification, Siam Review 55 (2013) 271–345

    2013

  10. [18]

    G.Winkler,etal., Integralrepresentationinthesetofsolutionsofageneralizedmomentproblem, MathematischeAnnalen246(1979)23–32

    1979

  11. [19]

    Winkler, Extreme points of moment sets, Mathematics of Operations Research 13 (1988) 581–587

    G. Winkler, Extreme points of moment sets, Mathematics of Operations Research 13 (1988) 581–587. R. Jin et al.:Preprint submitted to ElsevierPage 29 of 31 Table 10 Optimal bounds and relative deviations for the five-dimensional nonlinear problem. 𝑟 𝐾 𝑈RD of𝑈(%)𝐿RD of𝐿(%) 0 1 1...

    1988

  12. [20]

    Stenger, Optimal uncertainty quantification of a risk measurement from a computer code, Ph.D

    J. Stenger, Optimal uncertainty quantification of a risk measurement from a computer code, Ph.D. thesis, Université Paul Sabatier-Toulouse III, 2020

    2020

  13. [21]

    Kamga, B

    P.-H. Kamga, B. Li, M. McKerns, L. Nguyen, M. Ortiz, H. Owhadi, T. Sullivan, Optimal uncertainty quantification with model uncertainty and legacy data, Journal of the Mechanics and Physics of Solids 72 (2014) 1–19

    2014

  14. [22]

    Sun, Uncertainty quantification of material properties in ballistic impact of magnesium alloys, Materials 15 (2022) 6961

    X. Sun, Uncertainty quantification of material properties in ballistic impact of magnesium alloys, Materials 15 (2022) 6961

    2022

  15. [23]

    Balzani, T

    D. Balzani, T. Schmidt, M. Ortiz, Method for the quantification of rupture probability in soft collagenous tissues, International journal for numerical methods in biomedical engineering 33 (2017) e02781

    2017

  16. [24]

    Miska, D

    N. Miska, D. Balzani, Method for the analysis of epistemic and aleatory uncertainties for a reliable evaluation of failure of engineering structures, International Journal for Uncertainty Quantification 12 (2022)

    2022

  17. [25]

    Miska, D

    N. Miska, D. Balzani, Reliability-based design optimization incorporating extended optimal uncertainty quantification, Probabilistic Engineering Mechanics 80 (2025) 103755

    2025

  18. [26]

    J.Stenger,F.Gamboa,M.Keller, Optimizationofquasi-convexfunctionoverproductmeasuresets, SIAMJournalonOptimization31(2021) 425–447

    2021

  19. [27]

    D. Wang, H. Qiu, L. Gao, C. Jiang, A subdomain uncertainty-guided kriging method with optimized feasibility metric for time-dependent reliability analysis, Reliability Engineering & System Safety 243 (2024) 109839

    2024

  20. [28]

    Y. Zhao, J. Yang, M. G. Faes, S. Bi, Y. Wang, The sub-interval similarity: A general uncertainty quantification metric for both stochastic and interval model updating, Mechanical Systems and Signal Processing 178 (2022) 109319

    2022

  21. [29]

    J. Liu, X. Sun, K. Li, C. Jiang, X. Han, A probability density function discretization and approximation method for the dynamic load identification of stochastic structures, Journal of Sound and Vibration 357 (2015) 74–94. R. Jin et al.:Preprint submitted to ElsevierPage 30 of 31

    2015

  22. [30]

    Jiang, X

    C. Jiang, X. Han, G. Liu, Optimization of structures with uncertain constraints based on convex model and satisfaction degree of interval, Computer Methods in Applied Mechanics and Engineering 196 (2007) 4791–4800

    2007

  23. [31]

    X. Yin, W. Chen, A. To, C. McVeigh, W. K. Liu, Statistical volume element method for predicting microstructure–constitutive property relations, Computer methods in applied mechanics and engineering 197 (2008) 3516–3529

    2008

  24. [32]

    R.Ahlfeld,B.Belkouchi,F.Montomoli,Samba:Sparseapproximationofmoment-basedarbitrarypolynomialchaos,JournalofComputational Physics 320 (2016) 1–16

    2016

  25. [33]

    K.Sepahvand,S.Marburg, Identificationofcompositeuncertainmaterialparametersfromexperimentalmodaldata, ProbabilisticEngineering Mechanics 37 (2014) 148–153

    2014

  26. [34]

    Stenger, F

    J. Stenger, F. Gamboa, M. Keller, B. Iooss, Optimal uncertainty quantification of a risk measurement from a thermal-hydraulic code using canonical moments, International Journal for Uncertainty Quantification 10 (2020)

    2020

  27. [35]

    McKerns, H

    M. McKerns, H. Owhadi, C. Scovel, T. J. Sullivan, M. Ortiz, The optimal uncertainty algorithm in the mystic framework, arXiv preprint arXiv:1202.1055 (2012)

    2012 arXiv

  28. [36]

    Storn, K

    R. Storn, K. Price, Differential evolution–a simple and efficient heuristic for global optimization over continuous spaces, Journal of global optimization 11 (1997) 341–359

    1997

  29. [37]

    X. Sun, B. Liu, A learning-based optimal uncertainty quantification method and its application to ballistic impact problems, Mechanics of Materials 184 (2023) 104727

    2023

  30. [38]

    Dette, W

    H. Dette, W. J. Studden, The theory of canonical moments with applications in statistics, probability, and analysis, John Wiley & Sons, 1997

    1997

  31. [39]

    P. Henrici, Applied and computational complex analysis, Volume 3: Discrete Fourier analysis, Cauchy integrals, construction of conformal maps, univalent functions, volume 3, John Wiley & Sons, 1993

    1993

  32. [40]

    Gautschi, Orthogonal polynomials: computation and approximation, OUP Oxford, 2004

    W. Gautschi, Orthogonal polynomials: computation and approximation, OUP Oxford, 2004

    2004

  33. [41]

    G. H. Golub, J. H. Welsch, Calculation of gauss quadrature rules, Mathematics of computation 23 (1969) 221–230

    1969

  34. [42]

    C. P. Robert, G. Casella, G. Casella, Monte Carlo statistical methods, volume 2, Springer, 1999

    1999

  35. [43]

    R. Y. Rubinstein, D. P. Kroese, Simulation and the Monte Carlo method, John Wiley & Sons, 2016

    2016

  36. [44]

    Carlo, Monte carlo methods in financial engineering, 2001

    Q.-M. Carlo, Monte carlo methods in financial engineering, 2001

    2001

  37. [45]

    Rossetti, Simulation modeling using the kotlin simulation library (ksl), under the Creative Commons Attribution-Noncommercial-No Derivatives 4.0 International License (2023)

    M. Rossetti, Simulation modeling using the kotlin simulation library (ksl), under the Creative Commons Attribution-Noncommercial-No Derivatives 4.0 International License (2023)

    2023

  38. [46]

    K. V. Price, R. M. Storn, J. A. Lampinen, Differential evolution: a practical approach to global optimization, Springer, 2005

    2005

  39. [47]

    G. R. Johnson, A constitutive model and data for materials subjected to large strains, high strain rates, and high temperatures, Proc. 7th Inf. Sympo. Ballistics (1983) 541–547

    1983

  40. [48]

    Hasenpouth, Tensile high strain rate behavior of AZ31B magnesium alloy sheet, Master’s thesis, University of Waterloo, 2010

    D. Hasenpouth, Tensile high strain rate behavior of AZ31B magnesium alloy sheet, Master’s thesis, University of Waterloo, 2010

    2010

  41. [49]

    Grüneisen, Theorie des festen zustandes einatomiger elemente, Annalen der Physik 344 (1912) 257–306

    E. Grüneisen, Theorie des festen zustandes einatomiger elemente, Annalen der Physik 344 (1912) 257–306

    1912

  42. [50]

    G. I. Taylor, H. Quinney, The latent energy remaining in a metal after cold working, Proceedings of the Royal Society of London. Series A, Containing Papers of a Mathematical and Physical Character 143 (1934) 307–326

    1934

  43. [51]

    S. Lee, H. J. Ham, S. Y. Kwon, S. W. Kim, C. M. Suh, Thermal conductivity of magnesium alloys in the temperature range from- 125 c to 400 c, International Journal of Thermophysics 34 (2013) 2343–2350

    2013

  44. [52]

    O. T. Kingstedt, J. T. Lloyd, On the conversion of plastic work to heat in mg alloy az31b for dislocation slip and twinning deformation, Mechanics of Materials 134 (2019) 176–184

    2019

  45. [53]

    J. Feng, P. Chen, Q. Zhou, K. Dai, E. An, Y. Yuan, Numerical simulation of explosive welding using smoothed particle hydrodynamics method, The International Journal of Multiphysics 11 (2017) 315–326

    2017

  46. [54]

    URL:https://ftp.lstc

    LS-DYNA Theory Manual, Livermore Software Technology, ANSYS Company, Livermore, California, 2024. URL:https://ftp.lstc. com/anonymous/outgoing/web/ls-dyna_manuals/DRAFT/DRAFT_Theory.pdf

    2024

  47. [55]

    Courant, K

    R. Courant, K. Friedrichs, H. Lewy, Über die partiellen differenzengleichungen der mathematischen physik, Mathematische annalen 100 (1928) 32–74

    1928

  48. [56]

    C. M. Bishop, N. M. Nasrabadi, Pattern recognition and machine learning, volume 4, Springer, 2006

    2006

  49. [57]

    G.Klambauer,T.Unterthiner,A.Mayr,S.Hochreiter, Self-normalizingneuralnetworks, Advancesinneuralinformationprocessingsystems 30 (2017)

    2017

  50. [58]

    D. P. Kingma, Adam: A method for stochastic optimization, arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  51. [59]

    Prechelt, Early stopping-but when?, in: Neural Networks: Tricks of the trade, Springer, 2002, pp

    L. Prechelt, Early stopping-but when?, in: Neural Networks: Tricks of the trade, Springer, 2002, pp. 55–69

    2002

  52. [60]

    X. Sun, T. Kirchdoerfer, M. Ortiz, Rigorous uncertainty quantification and design with uncertain material models, International Journal of Impact Engineering 136 (2020) 103418. R. Jin et al.:Preprint submitted to ElsevierPage 31 of 31

    2020

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.