Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Producing a complete nuclear data library for adjoint Monte Carlo simulations

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read Adjoint Monte Carlo neutron transport can be run with a complete library of adjoint nuclear data, covering every reaction in modern nuclear data evaluations, and matches forward simulations on a 100-case reciprocity test.

desk verdict Complete adjoint nuclear data library from ENDF/B-VIII.0, with a solid collision-physics validation; spatial transport remains untested. read the letter →

arxiv 2607.22192 v1 pith:YOVBGPML submitted 2026-07-24 physics.comp-ph

Paul Rovel (LTSD , ENPC) , Coline Larmier (CEA-DES (ex-DEN)) , Davide Mancusi (LTSD , SERMA) , Andrea Zoia (SERMA) This is my paper · ORCID
classification physics.comp-ph
keywords adjointMonteCarloneutrontransportnucleardatareciprocitytheoremDopplerbroadeningthermalscatteringlawunresolvedresonancerangeprocessing
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

This paper aims to make adjoint Monte Carlo simulation—where fictitious particles are born at the detector and travel backward to the source—practical for real shielding problems by producing the missing ingredient: a complete adjoint nuclear data library. The method converts every reaction channel in modern evaluated nuclear data (elastic, discrete and continuum inelastic, fission, Kalbach, thermal scattering laws, unresolved resonance range, and Doppler-broadened elastic scattering) into adjoint cross sections and adjoint outgoing-energy/angle laws. The key maneuver is to define these adjoint laws by folding the direct data against an arbitrary importance-spectrum prior g(E), and to preserve unbiasedness through a weight correction that uses the same tabulated distributions for sampling and for the correction. The authors generate a full adjoint library for 555 nuclides and validate it on an infinite-medium U-238/H-1 benchmark: for 100 source–detector pairs, adjoint and direct Monte Carlo responses agree within statistical uncertainty (three pairs above two standard deviations, none above three). If the approach holds, adjoint games can exploit source–detector reciprocity to speed up calculations where a large source illuminates a small detector.

What carries the argument

The central object is the non-normalized adjoint distribution f̃†(E′→E) = g(E) σ(E) ν(E) f(E→E′, μ), whose integral over E gives the adjoint cross section and whose normalized version is the adjoint scattering law. It is stored with a two-dimensional mesh of piecewise-linear (unit-based) interpolations; sampling is done by stochastic interpolation between bracketing incident-energy distributions and inverse-transform sampling of the cumulative. The load-bearing consistency requirement is that the density used in the weight correction is exactly the density of the distribution sampled, which is what makes arbitrary discretization, angular simplifications, and Monte Carlo estimates of the Dopp

What would settle it

Compute or measure the variance of the single-sample estimator for the Doppler kernel on a strongly resonant nuclide at thermal energy (e.g., U-238 at 600 K or a synthetic narrow resonance) and check whether the adjoint game's figure of merit degrades dramatically compared with the forward game; if the estimator variance diverges with resonance strength, the efficiency claim fails even though bias may be absent.

Watch

Extended reading notes

Core claim

The paper establishes that a fully general treatment of adjoint collision sampling is possible: for every reaction present in modern evaluated nuclear data, an adjoint cross section and an adjoint distribution law can be precomputed and tabulated so that the adjoint game remains unbiased and efficient. Adjoint data are defined from direct data by folding them against a user-chosen prior guess g(E) of the direct flux shape; because the weight correction in the sampling equation is evaluated with the same discretized density used for sampling, tabulation errors do not bias results—they only degrade variance. The same principle covers thermal Doppler broadening, where the direct kernel is estim

Load-bearing premise

The practical convergence of the whole scheme rests on the claim that one Gaussian sample of the Doppler-broadened elastic-scattering kernel, together with a rejection-count estimate of its normalization, is an unbiased estimator; this holds in expectation for any cross section, but if the target cross section is strongly resonant the estimator's variance can explode, and without a variance bound the adjoint game may lose the efficiency it is meant to provide.

Editorial extensions

If this is right

  • Adjoint Monte Carlo can be implemented with continuous-energy nuclear data covering all reaction types, not just multigroup approximations, including thermal scattering laws and the unresolved resonance range.
  • Any detector response computable by a forward simulation is also computable by the adjoint game through reciprocity, so shielding configurations with a large source and a small detector become natural targets for variance reduction.
  • A complete adjoint library can be produced from a modern evaluated data set within about 20 CPU-minutes and about 2 GB of storage, making on-the-fly adjoint sampling feasible in a production code.
  • Discretization error, isotropic angular approximations, and even stochastic estimation of the thermal kernel do not bias the game, provided the weight correction uses the same sampled density—so the method degrades gracefully under approximation.
  • The approach is a stepping stone to adjoint transport in a next-generation production Monte Carlo code.

Reading between the lines

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

  • If the method scales past infinite-medium benchmarks, the same reciprocity test could be run in a spatially heterogeneous geometry (e.g., layered shielding) to check whether the weight-correction sampling remains practical when adjoint flights cross material boundaries.
  • The paper leaves the variance of the single-sample Doppler-kernel estimator unexamined; for a strongly resonant thermal scatterer, that estimator is unbiased but could produce very large weights, so a variance analysis or an adaptive number of samples near resonances is a natural extension.
  • The arbitrary prior g(E) is implicitly a tuning lever: choosing it from a preliminary forward calculation, or iterating between forward and adjoint runs, could push the game toward the zero-variance limit the authors mention, at the cost of an extra production step.
  • Because the formalism is stated for general particle transport, the same adjoint-data preparation could extend to photon transport or coupled neutron-photon games, though the paper demonstrates only neutrons.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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. This paper develops a general methodology for producing adjoint nuclear data libraries for continuous-energy Monte Carlo transport. Building on the authors' prior adjoint-sampling framework (Ref. [12]), it formulates adjoint cross sections and adjoint distribution laws from ENDF data using a flux-shape prior g(E), and details processing schemes for elastic and inelastic scattering, fission, continuum and Kalbach-type reactions, Doppler-broadened elastic scattering, thermal scattering laws, and unresolved resonance range data. It also presents adaptive discretization algorithms for unidimensional and bidimensional tabulated data. The authors produce an adjoint library for ENDF-B/VIII.0 and validate the collision sampling by a reciprocity test on an infinite homogeneous 238U/1H mixture, comparing 100 direct/adjoint source-detector responses; 3 of 100 t-values exceed 2σ and none exceed 3σ. The paper concludes that spatial-heterogeneity and flight-kernel tests remain future work.

Significance. If the claims hold, the paper removes a major obstacle to practical continuous-energy adjoint Monte Carlo: the preparation of adjoint data covering the full diversity of modern nuclear data formats. The systematic treatment of TSL, URR, thermal broadening, and all ENDF reaction laws is a substantial contribution, and the production of a full ENDF-B/VIII.0 adjoint library is an impressive engineering result. The reciprocity test is a meaningful consistency check that goes beyond simple unit tests. The paper is also honest about the limits of its validation: the benchmark is energy-only, the adjoint runs are reported to be 20× slower than direct runs, and several approximations are acknowledged but not quantified. These limitations do not invalidate the method, but they materially affect the strength of the paper's broader claims.

major comments (3)
  1. [Section V.B/V.C and Section II.B.1] The validation is performed in an infinite homogeneous medium, so the problem 'only depends on the energy variable.' In this geometry the adjoint flight operator (Eq. (5)) reduces to a trivial case: for energy-conserving flights in a spatially constant medium, the weight correction Σt(r′,E′)/Σt(r,E) is identically 1, and boundary crossings are never exercised. The 100 reciprocity pairs therefore test the collision kernel and its energy dependence, but not the spatial transport components of the adjoint game. The Conclusions explicitly state that 'a detailed study of the behavior of adjoint Monte Carlo games in the case involving flights and spatially heterogeneous media is needed.' Since the stated goal is a full adjoint Monte Carlo capability for shielding applications, the present evidence does not yet support the general validity of the method for those problems. A spatially heterogen
  2. [Section III.H.2, Eqs. (72)-(83)] The on-the-fly evaluation of the Doppler-broadened scattering kernel relies on a single-sample Monte Carlo estimate of the integral in Eq. (77) and on a rejection-count estimator for C−1(v) in Eq. (83). The authors correctly argue that these estimators are unbiased in expectation, but no variance or convergence analysis is provided. For strongly resonant 0 K cross sections, the single-sample estimate can have very large fluctuations, which would not bias the adjoint game but could destroy the efficiency that motivates the method. This concern is concrete: Section V.C reports that the adjoint simulations took, on average, 20 times longer than the direct simulations to reach comparable statistical uncertainty. The manuscript should at least report empirical variances of these thermal estimators on the benchmark, or a theoretical bound, and discuss the implications for practical use.
  3. [Section IV.E and Section V.A] The discretization tolerances (10% for unidimensional, 20% for bidimensional data) and the 'leveler' mechanism are introduced to control storage size and variance, but no quantitative connection is made between these tolerances and the observed benchmark results. The weight-correction factor scales as 1/rd/o, where rd/o = fd/fo; in the 'leveled' regions, rd/o can be arbitrarily large, potentially producing very large particle weights. The paper states that the chosen settings 'have been found to work well on several test cases,' but no sensitivity study is provided. Since the central practical claim is that the discretized adjoint data preserve the variance-reduction properties of the idealized definitions in Eq. (12), the authors should either quantify the actual deviations of the produced data from the target tolerances or show that benchmark results are insensitive to the tolerance ch
minor comments (5)
  1. [Section III.H.2] The sentence 'we use the sampling' appears to contain a typo; it should probably read 'we use the sampling procedure' or 'we use the DBRC strategy.'
  2. [Appendix A, Eq. (84)] The expression for P(E→E′) has a somewhat unbalanced use of parentheses around the error functions; this makes the formula harder to read. A cleaner bracketing would help readers verify the signs for E<E′ and E>E′.
  3. [Section V.C, Table II] The 100 t-values are not independent, because for a given source the 10 detector responses are tallied in the same simulation and hence correlated. The statement that the t-values 'should follow a standard normal distribution' is therefore only an approximation. This does not affect the main conclusion, but it should be acknowledged.
  4. [Section IV.D] The note that the two-stage Douglas-Peucker application yields a final tolerance 'twice as large as the individual tolerance' is potentially confusing. It should be clarified whether each of the two stages is run with the stated tolerance, so that the total error bound is the sum, or whether the individual tolerances are halved to keep the final bound.
  5. [Section III.F] The choice to identify the fission spectrum χf with the spectrum at the highest incident energy when U is negative, and the arbitrary −10 MeV threshold for treating the spectrum as energy-dependent, are heuristic. The text explains the variance rationale, but it would be useful to state explicitly that this choice cannot cause bias because the weight correction uses the true direct distribution.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: adjoint data are deterministic transforms of ENDF data plus a user prior, and the reciprocity validation is an implementation self-consistency test rather than a fitted prediction.

full rationale

The paper's central derivation is self-contained: adjoint cross sections and distributions are defined by Eq. (12) as integrals of direct ENDF quantities (σ_i,j, ν_i,j, f_i,j) against a user-supplied prior g(E). No target response is fitted, and the prior g(E) is not inferred from the 100 reciprocity-test responses. The unbiasedness of the adjoint game is guaranteed by construction through the weight correction Eq. (11), and the paper explicitly states that the adjoint data 'can be chosen arbitrarily' provided the weight correction satisfies Eq. (11). Thus the numerical comparison with forward Monte Carlo is a verification of the implementation and of the discretization choices, not a demonstration that a fitted parameter reproduces its own input. The self-citations to Ref. [12] (general adjoint sampling strategy) and Ref. [29] (reciprocity test) are not load-bearing in a circular sense: the present paper restates the sampling equations and the statistical test is standard. The acknowledged limitation that the benchmark is an infinite homogeneous medium—'the problem only depends on the energy variable'—and the conclusion that 'a detailed study of the behavior of adjoint Monte Carlo games in the case involving flights and spatially heterogeneous media is needed' further show that the validation is appropriately scoped. The single-sample thermal-estimator variance issue is an efficiency concern, not a circularity: the estimators are argued to be unbiased and independent, and the paper does not claim zero variance. Overall, no circular step reduces the claimed results to their inputs.

Assumptions & free parameters 8 free parameters · 7 assumptions · 0 invented entities

The central result rests on standard transport theory plus the unproved-in-this-work unbiasedness of arbitrary adjoint data with weight corrections. No new physical entity is introduced. The free parameters are mostly algorithmic precision/variance levers rather than physics fits, except for the 20 kBT thermal cutoff, which is an admitted small bias.

free parameters (8)
  • Flux-shape prior g(E) = 1/E (with optional thermal Maxwellian at kBT = 0.1 eV in illustrations)
    Chosen by hand as a variance-shaping prior in Eq. (12); not fitted to benchmark responses. Affects efficiency, not unbiasedness.
  • Thermal truncation energy E_target,max = 20 kBT
    Arbitrary cutoff in Section III.H.3 for the thermal adjoint distribution; introduces a small bias estimated at ~1e-8 probability of truncation.
  • Discretization tolerances = 10% (1D), 20% (2D)
    User-selected precision/size trade-off in Section IV.E; the impact on adjoint-game variance is not quantitatively assessed.
  • Discretization algorithm settings = 300 points/decade (log) or 100 (linear); n_test = 7; max 1e5 points/decade; 200 functions/decade
    Heuristic algorithmic parameters in Section IV.B/IV.E that determine the discretized library size and fidelity.
  • Fission-independent-spectrum threshold U = -10 MeV
    Arbitrary threshold in Section III.F deciding when a fission spectrum is treated as independent of incident energy.
  • CM-frame angular integration points = 3 (A>10) or 5 (A<=10) mu_l points
    Numerical integration in Section III.G.2 for converting CM-frame densities to laboratory frame; approximation used to build generic energy-to-energy adjoint laws.
  • Singularity avoidance cutoff = mu_l in [-1, 0.98]
    Chosen in Section III.G.2 to avoid E'_cm = 0 singularities in the laboratory-frame density conversion.
  • Leveler cumulative threshold = 0.2% of cumulative
    Parameter in Section IV.E that assigns a minimum value to f_d in low-probability regions, changing the stored distribution without affecting unbiasedness.
assumptions (7)
  • standard math Reciprocity theorem and adjoint transport equation duality (R = <S, chi-dagger> = <psi, eta_psi>)
    Used throughout Section II as the mathematical basis for replacing direct with adjoint simulation; not proved in this paper.
  • domain assumption Any positive adjoint data sigma-dagger, f-dagger with weight correction nu-dagger satisfying Eq. (11) yields an unbiased adjoint Monte Carlo game
    Stated in Section II.B.2/II.C, following Refs [12, 13]; underpins the freedom to prepare adjoint data via Eq. (12) and tolerate discretization error.
  • domain assumption ENDF-B/VIII.0 ACE-formatted data, parsed via NJOY and ALEXANDRIA, correctly represents the physics
    The input to library production in Section V.A; no independent validation of the source data is performed.
  • domain assumption Target nucleus velocity obeys a Maxwellian (SVT) at temperature T when TSL data are absent
    Used in Section III.H to derive the Doppler-broadened scattering kernel; standard but an approximation for real materials.
  • domain assumption Continuous TSL interpolation scheme (Eqs. 95-99) is the correct representation of inelastic TSL distributions
    Adopted from ACE/ENDF practice in Section III.I; the paper builds the adjoint TSL on this interpolation.
  • ad hoc to paper The single-Gaussian-sample estimator of integral (77) and the rejection-count estimator (83) are unbiased and independent
    Introduced in Section III.H.2 to make the thermal weight correction tractable; their high variance in resonant regimes is not analyzed.
  • ad hoc to paper Truncating thermal adjoint distribution at E_target,max = 20 kBT introduces negligible bias (~1e-8 probability)
    Arbitrary cutoff in Section III.H.3; creates a region where adjoint support is absent while direct support is non-zero.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Producing a complete nuclear data library for adjoint Monte Carlo simulations." pith.science (2026). https://pith.science/paper/YOVBGPML

@misc{pith2026260722192,
  author       = {Pith},
  title        = {Pith review of: Producing a complete nuclear data library for adjoint Monte Carlo simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YOVBGPML}},
  note         = {Machine review of arXiv:2607.22192}
}
read the original abstract

Radiation shielding applications related to reactor design typically involve situations where the source region (the core) is much larger than the detector region (a dosimeter). In such cases, the efficiency of Monte Carlo simulation might be significantly increased by solving the adjoint transport equation: adjoint particles are born from the detector, undergo adjoint ('reversed') flights and collisions, and accumulate their tallies in the source region. A key prerequisite to sample the adjoint collision events is the preparation of adjoint nuclear data. In this work, we propose a general method able to handle the diversity of nuclear reactions available in modern neutron data evaluations, and ultimately create a full adjoint nuclear data library. This is a stepping stone in view of implementing adjoint sampling schemes in TRIPOLI-5 __ , the next-generation Monte Carlo code developed by CEA and ASNR. We validate this strategy based on a relevant continuous energy benchmark configuration involving mixtures of heavy and light nuclides, and we compare our results to those obtained by standard forward Monte Carlo simulations.

Figures

Figures reproduced from arXiv: 2607.22192 by the authors.

Figure 1
Figure 1. Bidimensional interpolation scheme for the storage of [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. The bidimensional unit-based interpolation scheme to preserve the distribution edges. After [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. Unit-based interpolation scheme. See text for detailed procedure. Black dots represent the [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: Adjoint distribution law for the elastic scattering of [PITH_FULL_IMAGE:figures/full_fig_p021_4.png]
Figure 5
Figure 5. Figure 5: Adjoint cross sections σ † (E′ ) for elastic scattering of 238U (left) and 1H (right). As the formula for the adjoint cross section Eq. (12) effectively acts as a rolling average over the range of accessible energies, the adjoint cross section features smoothed resonan…
Figure 6
Figure 6. Figure 6: Adjoint nuclear data for the inelastic scattering on the first discrete level ( [PITH_FULL_IMAGE:figures/full_fig_p025_6.png]
Figure 7
Figure 7. Figure 7: Adjoint nuclear data of the fission reaction of [PITH_FULL_IMAGE:figures/full_fig_p028_7.png]
Figure 8
Figure 8. Figure 8: Procedure to obtain the bounds of the adjoint distribution. The distribution is shown in [PITH_FULL_IMAGE:figures/full_fig_p030_8.png]
Figure 9
Figure 9. Figure 9: Adjoint nuclear data of the (n, 2n) reaction of 238U. On the left, the adjoint cross section σ † (E′ ). On the right, the non-normalized adjoint distribution law ˜f † . This discretized distribution contains 1020 data points, split into 73 unidimensional distributions,…
Figure 10
Figure 10. Figure 10: Adjoint nuclear data of the thermal part (SVT approximation) of the elastic scattering [PITH_FULL_IMAGE:figures/full_fig_p044_10.png]
Figure 11
Figure 11. Figure 11: Procedure to obtain the density with the interpolation of TSL laws. One finds the closest [PITH_FULL_IMAGE:figures/full_fig_p047_11.png]
Figure 12
Figure 12. Figure 12: Adjoint nuclear data of the TSL of 238U (in UO2 at 1200 K). Data produced using Eq. (12) with a g(E) featuring a 1/E neutron slowdown shape and a thermal Maxwellian shape at kBT = 0.1 eV (≈ 1200 K). On the left, the adjoint cross section σ † (E′ ) of the different TSL…
Figure 13
Figure 13. Figure 13: Adjoint URR state distribution of the elastic scattering of [PITH_FULL_IMAGE:figures/full_fig_p050_13.png]
Figure 14
Figure 14. Figure 14: Discretization procedure using the additive algorithm. The dotted line represents the [PITH_FULL_IMAGE:figures/full_fig_p054_14.png]
Figure 15
Figure 15. Figure 15: Fail cases of the additive discretization algorithm where a segment is incorrectly validated. [PITH_FULL_IMAGE:figures/full_fig_p055_15.png]
Figure 16
Figure 16. Figure 16: Discretization procedure using the Douglas-Peucker algorithm. The dotted line repre [PITH_FULL_IMAGE:figures/full_fig_p057_16.png]
Figure 17
Figure 17. Figure 17: 2D discretization mesh D, for the energy-to-energy adjoint law ˜f † prepared for the adjoint SVT reaction of 238U. The discretized function is presented in the right part of [PITH_FULL_IMAGE:figures/full_fig_p061_17.png]
Figure 18
Figure 18. Figure 18: Collided importance ψ † relative to a Dirac detector at E∗ = 0.052 eV for a test problem consisting of 238U and 1H. See text for precise problem definition. The collision importance ψ † , which is proportional to the probability of a neutron scoring in the detector, h…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 10 canonical work pages

  1. [12]

    Rovel , C

    P. Rovel , C. Larmier , D. Mancusi , and A. Zoia , A general Monte Carlo sampling strategy for the adjoint Boltzmann equation in continuous energy, Nuclear Science and Engineering (2025); 10.1080/00295639.2025.2597634

    arXiv 2025

  2. [1]

    C. W. Maynard , An Application of the Reciprocity Theorem to the Acceleration of Monte Carlo Calculations , Nuclear Science and Engineering, 10, 2, 97 (1961); 10.13182/NSE61-A25945 ., ://doi.org/10.13182/NSE61-A25945

    work page doi:10.13182/nse61-a25945 1961

  3. [3]

    Lux and L

    I. Lux and L. Koblinger , Monte Carlo Particle Transport Methods : Neutron and Photon Calculations , CRC Press (2018) ://www.taylorfrancis.com/books/9781351083287

    arXiv 2018

  4. [4]

    De Matteis , Phenomenological interpretation of the adjoint neutron transport equation, Meccanica, 9, 3, 162 (1974); 10.1007/BF02317707 ., ://doi.org/10.1007/BF02317707

    A. De Matteis , Phenomenological interpretation of the adjoint neutron transport equation, Meccanica, 9, 3, 162 (1974); 10.1007/BF02317707 ., ://doi.org/10.1007/BF02317707

    work page doi:10.1007/bf02317707 1974

  5. [5]

    J. E. Hoogenboom , Adjoint Monte Carlo methods in neutron transport calculations, PhD Thesis , Delft University Press (1977)

    1977

  6. [6]

    J. E. Hoogenboom , A Practical Adjoint Monte Carlo Technique for Fixed - Source and Eigenfunction Neutron Transport Problems , Nuclear Science and Engineering, 79, 4, 357 (1981); 10.13182/NSE81-A21387 ., ://doi.org/10.13182/NSE81-A21387

    work page doi:10.13182/nse81-a21387 1981

  7. [7]

    J. A. Kulesza , T. R. Adams , J. C. Armstrong , S. R. Bolding , F. B. Brown , J. S. Bull , T. P. Burke , A. R. Clark , R. A. Forster, III , J. F. Giron , A. S. Grieve , C. J. Josey , R. L. Martz , G. W. McKinney , E. J. Pearson , M. E. Rising , C. J. Solomon, Jr. , S. Swaminarayan , T. J. Trahan , C. A. Weaver , S. C. Wilson , and A. J. Zukaitis , MCNP Co...

    arXiv 2024

  8. [8]

    D. A. Brown , ENDF-6 Formats Manual - Data Formats and Procedures for the Evaluated Nuclear Data Files ENDF/B-VI, ENDF/B-VII and ENDF/B-VIII, BNL--224854-2023-INRE, Brookhaven National Laboratory (BNL), Upton, NY (United States) (2023); 10.2172/2007538 ., ://www.osti.gov/biblio/2007538

    arXiv 2023

Show all 27 references
  1. [9]

    M. Grimstone , Extension of the MCBEND Monte Carlo code to perform adjoint calculations using point energy data, ANS Radiation Protection and Shielding Division Topical Conference, Nashville, 143--150 (1998)

    1998

  2. [10]

    J. E. Hoogenboom , Methodology of continuous-energy adjoint Monte Carlo for neutron, photon, and coupled neutron-photon transport, Nuclear science and engineering, 143, 2, 99 (2003)

    2003

  3. [11]

    C. M. Diop , O. Petit , C. Jouanne , and M. Coste-Delclaux , Adjoint Monte Carlo neutron transport using cross-section probability table representation, Annals of Nuclear Energy, 37, 9, 1186 (2010); https://doi.org/10.1016/j.anucene.2010.04.018

    2010 doi

  4. [13]

    A. D. Matteis and R. Simonini , A New Monte Carlo Approach to the Adjoint Boltzmann Equation , Nuclear Science and Engineering, 65, 1, 93 (1978); 10.13182/NSE78-A27129

    1978 doi

  5. [15]

    L. L. Carter and N. J. McCormick , Coupled Sampling with the Monte Carlo Method in Neutron Transport Calculations , Nuclear Science and Engineering, 39, 3, 296 (1970); 10.13182/NSE70-A19991 ., ://doi.org/10.13182/NSE70-A19991

    1970 doi

  6. [16]

    Saracco , S

    P. Saracco , S. Dulla , and P. Ravetto , The adjoint neutron transport equation and the statistical approach for its solution, The European Physical Journal Plus, 131, 11, 412 (2016); 10.1140/epjp/i2016-16412-0 ., ://link.springer.com/10.1140/epjp/i2016-16412-0

    2016 doi

  7. [17]

    J. L. Conlin and P. Romano , A compact ENDF ( ACE ) format specification, , Los Alamos National Laboratory (LANL), Los Alamos, NM (United States) (2019) ://www.osti.gov/servlets/purl/1561065

    2019

  8. [18]

    G. I. Bell and S. Glasstone , Nuclear Reactor Theory , TID--25606, United States Atomic Energy Commission (1970) ://www.osti.gov/biblio/4074688

    1970

  9. [19]

    Zoia , E

    A. Zoia , E. Brun , C. Jouanne , and F. Malvagi , Doppler broadening of neutron elastic scattering kernel in Tripoli -4®, Annals of Nuclear Energy, 54, 218 (2013); 10.1016/j.anucene.2012.11.023 ., ://www.sciencedirect.com/science/article/pii/S0306454912004689

    2013 doi

  10. [20]

    W. Rothenstein , Neutron scattering kernels in pronounced resonances for stochastic Doppler effect calculations, Annals of Nuclear Energy, 23, 4, 441 (1996); 10.1016/0306-4549(95)00109-3 ., ://www.sciencedirect.com/science/article/pii/0306454995001093

    1996

  11. [21]

    T. E. Booth , Unbiased Monte Carlo Estimation of the Reciprocal of an Integral , Nuclear Science and Engineering, 156, 3, 403 (2007); 10.13182/NSE07-A2707 ., ://www.tandfonline.com/doi/full/10.13182/NSE07-A2707

    2007 doi

  12. [22]

    NIST Digital Library of Mathematical Functions , https://dlmf.nist.gov/, Release 1.2.7 of 2026-06-15 ://dlmf.nist.gov/, f. W. J. Olver, A. B. Olde Daalhuis , D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, B. V. Saunders, H. S. Cohl, and M. A. McClain, eds

    2026

  13. [23]

    Hörmann and J

    W. Hörmann and J. Leydold , Generating generalized inverse Gaussian random variates, Statistics and Computing, 24, 4, 547 (2014); 10.1007/s11222-013-9387-3 ., ://doi.org/10.1007/s11222-013-9387-3

    2014 doi

  14. [24]

    Sublet , P

    J.-C. Sublet , P. Ribon , M. Coste-Delclaux , and C. a. l. A. CEA (France) , CALENDF -2010: user manual, CEA-R--6277, Commissariat a l'energie Atomique - CEA (France) (2011) ://inis.iaea.org/records/xk7cg-jne31

    2010

  15. [25]

    R. E. Macfarlane et al. , The NJOY Nuclear Data Processing System , Version 2016, LA-UR--17-20093, Los Alamos National Lab. (LANL), Los Alamos, NM (United States) (2017)

    2016

  16. [26]

    D. H. Douglas and T. K. Peucker , Algorithms for the reduction of the number of points required to represent a digitized line or its caricature, Cartographica, 10, 2, 112 (1973); 10.3138/FM57-6770-U75U-7727 ., ://utppublishing.com/doi/abs/10.3138/fm57-6770-u75u-7727

    1973 doi

  17. [27]

    J. E. Hershberger and J. Snoeyink , Speeding up the Douglas - Peucker line-simplification algorithm, , University of British Columbia, Department of Computer Science Vancouver, BC (1992) ://www.uni-trier.de/fileadmin/fb4/prof/INF/DEA/Seminar0708/Hershberger-Snoeyink1.pdf

    1992

  18. [28]

    Mancusi et al

    D. Mancusi et al. , Overview of TRIPOLI -5, a Monte Carlo code for HPC , EPJ N - Nuclear Sciences & Technologies, 10, 26 (2024); 10.1051/epjn/2024028

    2024

  19. [29]

    Rovel , C

    P. Rovel , C. Larmier , D. Mancusi , and A. Zoia , Preparing Nuclear Data for Adjoint Monte Carlo Calculations , Proceedings of PHYSOR 2026 International Conference , Politecnico di Torino, Torino, Italy (2026); 10.5281/zenodo.20803399 ., ://zenodo.org/records/20803399

    2026

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.