pith. sign in

arxiv: 1907.02208 · v1 · pith:2FADQFBInew · submitted 2019-07-02 · 📡 eess.SY · cs.SY

Surrogate model approach for investigating the stability of a friction-induced oscillator of Duffing's type

Jan N. Fuhg , Amelie Fau This is my paper

Pith reviewed 2026-05-25 10:45 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords surrogate modelkrigingDuffing oscillatorfriction-induced vibrationLyapunov exponentstability analysisparametric studyadaptive sampling
0
0 comments X

The pith

An adaptive kriging classification method called MiVor maps instability domains in a friction-induced Duffing oscillator by classifying the largest Lyapunov exponent.

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

The paper proposes a surrogate strategy that combines Gaussian-process kriging with an adaptive classification algorithm, MiVor, to explore the parameter space of a Duffing oscillator that includes an elasto-plastic friction force. It treats the largest Lyapunov exponent, obtained from a perturbation calculation, as the primary label for non-regular motion and also tracks sticking time. The resulting surrogate identifies regions of instability without exhaustive sampling of the full parameter domain. A sympathetic reader would care because direct parametric sweeps become prohibitive once the mechanical response is disrupted by friction and the dimension of the parameter space grows.

Core claim

The paper establishes that the MiVor adaptive kriging strategy for classification is highly proficient at detecting instabilities across the parametric space of the friction-induced Duffing oscillator and remains effective for complex response surfaces even in multi-dimensional parametric domains.

What carries the argument

MiVor, an adaptive kriging strategy for classification that uses Gaussian processes to label parameter points according to whether the estimated largest Lyapunov exponent indicates non-regular behavior.

If this is right

  • Instability domains can be located with far fewer full-model evaluations than a dense grid search requires.
  • Sticking time becomes a practical quantity of interest that can itself be modeled by kriging.
  • The same classification workflow extends directly to response surfaces that are non-smooth or high-dimensional.
  • Perturbation-based Lyapunov estimation supplies a usable label for machine-learning classification in this class of systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach could be tested on other discontinuous oscillators whose stability boundaries are similarly expensive to trace by brute force.
  • If the Lyapunov label remains accurate, the surrogate could be reused as a fast constraint inside optimization loops for friction-tuned devices.
  • The method implicitly suggests that classification surrogates may outperform regression surrogates when the goal is only to delineate stable versus unstable regions rather than to predict exact amplitudes.

Load-bearing premise

The largest Lyapunov exponent estimated by a perturbation method is an efficient and reliable indicator of non-regular behavior for this friction-induced oscillator.

What would settle it

A direct long-time integration of the oscillator equations at a parameter point where the surrogate predicts a negative Lyapunov exponent yet the simulated trajectory exhibits sustained non-periodic motion, or the reverse.

read the original abstract

Parametric studies for dynamic systems are of high interest to detect instability domains. This prediction can be demanding as it requires a refined exploration of the parametric space due to the disrupted mechanical behavior. In this paper, an efficient surrogate strategy is proposed to investigate the behavior of an oscillator of Duffing's type in combination with an elasto-plastic friction force model. Relevant quantities of interest are discussed. Sticking time is considered using a machine learning technique based on Gaussian processes called kriging. The largest Lyapunov exponent is proposed as an efficient indicator of non-regular behavior. This indicator is estimated using a perturbation method. A dedicated adaptive kriging strategy for classification called MiVor is utilized and appears to be highly proficient in order to detect instabilities over the parametric space and can furthermore be used for complex response surfaces in multi-dimensional parametric domains.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes an efficient surrogate-model strategy for parametric stability analysis of a Duffing oscillator augmented by an elasto-plastic friction law. Sticking time is modeled via Gaussian-process kriging; non-regular behavior is flagged by the sign of the largest Lyapunov exponent (LLE) obtained from a perturbation method. A dedicated adaptive classification kriging procedure (MiVor) is applied to map instability domains over the parameter space and is asserted to be highly proficient even for complex, multi-dimensional response surfaces.

Significance. If the claimed proficiency of MiVor is substantiated by quantitative validation and if the perturbation-based LLE remains a faithful classifier across stick-slip switches, the method would offer a practical tool for exploring high-dimensional parameter spaces in non-smooth mechanical oscillators without exhaustive time integration. The combination of classification-oriented adaptive kriging with an independent dynamical indicator is a potentially reusable contribution for surrogate-based bifurcation studies.

major comments (2)
  1. [Abstract and §4 (Results)] The central claim that MiVor 'appears to be highly proficient' (Abstract) rests on unshown numerical evidence: no classification error rates, confusion matrices, or comparison against direct LLE sampling are supplied anywhere in the manuscript. Without these metrics the asserted advantage over standard grid or Monte-Carlo exploration cannot be evaluated.
  2. [§3.2] §3.2 (LLE computation): the perturbation method for the largest Lyapunov exponent is applied to a vector field that is only piecewise Lipschitz because of the elasto-plastic friction law and the associated velocity discontinuities. No verification is given that the sign of the computed LLE correlates with other diagnostics (boundedness of long trajectories, Poincaré sections, or power spectra) across the stick-slip transitions; this assumption is load-bearing for the entire MiVor classification.
minor comments (2)
  1. [§2] Notation for the friction parameters and the perturbation vector should be introduced once and used consistently; several symbols appear without prior definition in §2.
  2. [Figures 4-7] Figure captions should state the number of training points used for each MiVor run and the final misclassification rate on a held-out test set.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help improve the clarity and rigor of the manuscript. We address each major comment below and will incorporate the suggested additions in the revised version.

read point-by-point responses

  1. Referee: [Abstract and §4 (Results)] The central claim that MiVor 'appears to be highly proficient' (Abstract) rests on unshown numerical evidence: no classification error rates, confusion matrices, or comparison against direct LLE sampling are supplied anywhere in the manuscript. Without these metrics the asserted advantage over standard grid or Monte-Carlo exploration cannot be evaluated.

    Authors: We agree that quantitative validation metrics are essential to support the claim. In the revised manuscript we will add classification error rates, confusion matrices, and direct comparisons of MiVor predictions against exhaustive LLE sampling on a validation set. These additions will quantify the proficiency and allow explicit evaluation against grid/Monte-Carlo approaches. revision: yes

  2. Referee: [§3.2] §3.2 (LLE computation): the perturbation method for the largest Lyapunov exponent is applied to a vector field that is only piecewise Lipschitz because of the elasto-plastic friction law and the associated velocity discontinuities. No verification is given that the sign of the computed LLE correlates with other diagnostics (boundedness of long trajectories, Poincaré sections, or power spectra) across the stick-slip transitions; this assumption is load-bearing for the entire MiVor classification.

    Authors: We acknowledge the concern regarding the piecewise-Lipschitz nature of the system. In the revised manuscript we will include additional verification: for representative parameter points on both sides of stick-slip transitions we will compare the sign of the perturbation-based LLE against long-term boundedness of trajectories, Poincaré sections, and spectral content. This will confirm that the LLE sign remains a reliable classifier across the discontinuities. revision: yes

Circularity Check

0 steps flagged

No significant circularity; surrogate trained on independent simulations

full rationale

The paper trains adaptive kriging (MiVor) and uses largest Lyapunov exponent (via perturbation) as an indicator of non-regular behavior. Both the training data and the LLE computation are obtained from external numerical simulations of the oscillator; neither quantity is defined in terms of the other or reduced to a fitted parameter renamed as a prediction. No self-citation chains, uniqueness theorems, or ansatzes imported from prior author work appear as load-bearing steps in the described methodology. The derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on standard dynamical-systems assumptions about Lyapunov exponents and the representativeness of the chosen friction model; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption Largest Lyapunov exponent reliably flags non-regular behavior
    Abstract proposes it as the indicator without further justification.

pith-pipeline@v0.9.0 · 5673 in / 1004 out tokens · 41757 ms · 2026-05-25T10:45:06.535686+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages · 1 internal anchor

  1. [1]

    Stick and slip

    E Rabinowicz. Stick and slip. Scientific American, 194(5):109–119, 1956

    work page 1956

  2. [2]

    Braking 2004: Vehicle Braking and Chassis Control, volume 6

    D Barton and A Blackwood. Braking 2004: Vehicle Braking and Chassis Control, volume 6. John Wiley & Sons, 2004

    work page 2004

  3. [3]

    Investigation of stick-slip vi- bration in a commercial vehicle brake assembly

    N Ashraf, D Bryant, and JD Fieldhouse. Investigation of stick-slip vi- bration in a commercial vehicle brake assembly. International Journal of Acoustics & Vibration, 22(3), 2017

    work page 2017

  4. [4]

    The reduction of stick-slip fric- tion in hydraulic actuators

    William S Owen and Elizabeth A Croft. The reduction of stick-slip fric- tion in hydraulic actuators. IEEE/ASME transactions on mechatronics , 8(3):362–371, 2003

    work page 2003

  5. [5]

    Comparisons of draft gear damping mechanisms

    Q Wu, S Luo, Qu T, and X Yang. Comparisons of draft gear damping mechanisms. Vehicle System Dynamics, 55(4):501–516, 2017. 30 Jan N. Fuhg, Am´ elie Fau

    work page 2017

  6. [6]

    Structural stiffness, dry friction coefficient, and equivalent viscous damping in a bump-type foil gas bearing

    D Rubio and L San Andres. Structural stiffness, dry friction coefficient, and equivalent viscous damping in a bump-type foil gas bearing. Journal of engineering for gas turbines and power , 129(2):494–502, 2007

    work page 2007

  7. [7]

    Stick-slip vibrations and chaos

    P Stelter. Stick-slip vibrations and chaos. Phil. Trans. R. Soc. Lond. A , 332(1624):89–105, 1990

    work page 1990

  8. [8]

    Dynamics of oscillators with impact and friction

    N Hinrichs, M Oestreich, and K Popp. Dynamics of oscillators with impact and friction. Chaos, Solitons & Fractals , 8(4):535–558, 1997

    work page 1997

  9. [9]

    Dynamics of a simple damped oscillator undergoing stick-slip vibrations

    U Galvanetto and S Bishop. Dynamics of a simple damped oscillator undergoing stick-slip vibrations. Meccanica, 34(5):337–347, 1999

    work page 1999

  10. [10]

    Two FEM approaches for the prediction and quantification of stick-slip phenomena on rubber- metal sliding contacts

    M Jim´ enez, J Bielsa, R Rodr´ ıguez, and C Bernad. Two FEM approaches for the prediction and quantification of stick-slip phenomena on rubber- metal sliding contacts. In IUTAM Symposium on Computational Methods in Contact Mechanics , pages 291–309. Springer, 2007

    work page 2007

  11. [11]

    Analytical approximations for stick–slip amplitudes and frequency of duffing oscillator

    K Devarajan and B Balaram. Analytical approximations for stick–slip amplitudes and frequency of duffing oscillator. Journal of Computational and Nonlinear Dynamics , 12(4):044501, 2017

    work page 2017

  12. [12]

    Discontinuity induced bifurcations in nonlinear systems

    B Santhosh, S Narayanan, and C Padmanabhan. Discontinuity induced bifurcations in nonlinear systems. Procedia IUTAM, 19:219–227, 2016

    work page 2016

  13. [13]

    Stick-slip chaos detection in coupled oscillators with friction

    J Awrejcewicz and D Sendkowski. Stick-slip chaos detection in coupled oscillators with friction. International Journal of Solids and Structures , 42(21-22):5669–5682, 2005

    work page 2005

  14. [14]

    Spectrum of Lyapunov exponents in non-smooth systems evaluated using orthogonal perturbation vectors

    M Balcerzak, A Dabrowski, A Stefa´ nski, and J Wojewoda. Spectrum of Lyapunov exponents in non-smooth systems evaluated using orthogonal perturbation vectors. In MATEC Web of Conferences , volume 148, page 10003. EDP Sciences, 2018

    work page 2018

  15. [15]

    Numerical analysis of the friction-induced oscillator of Duffing’s type with modified LuGre friction model

    D Pikunov and A Stefanski. Numerical analysis of the friction-induced oscillator of Duffing’s type with modified LuGre friction model. Journal of Sound and Vibration , 440:23–33, 2019

    work page 2019

  16. [16]

    Kolmogorov entropy and numerical experiments

    G Benetti, L Galgani, and J Strelcyn. Kolmogorov entropy and numerical experiments. Physical Review A, 14(6):2338, 1976

    work page 1976

  17. [17]

    Characterization of strange attractors

    P Grassberger and I Procaccia. Characterization of strange attractors. Physical Review Letters, 50(5):346, 1983

    work page 1983

  18. [18]

    Stick-mode duration of a dry-friction oscillator with an uncertain model

    R Lima and R Sampaio. Stick-mode duration of a dry-friction oscillator with an uncertain model. Journal of Sound and Vibration , 353:259–271, 2015

    work page 2015

  19. [19]

    Discrete chaos-i: Theory

    L Kocarev, J Szczepanski, J Amig´ o, and I Tomovski. Discrete chaos-i: Theory. IEEE Transactions on Circuits and Systems I: Regular Papers , 53(6):1300–1309, 2006

    work page 2006

  20. [20]

    Quantifying chaos by various computational methods

    J Awrejcewicz, A Krysko, N Erofeev, V Dobriyan, M Barulina, and V Krysko. Quantifying chaos by various computational methods. part 1: Simple systems. Entropy, 20(3):175, 2018

    work page 2018

  21. [21]

    Regression and kriging metamodels with their experimental designs in simulation: a review

    J Kleijnen. Regression and kriging metamodels with their experimental designs in simulation: a review. European Journal of Operational Research, 256(1):1–16, 2017

    work page 2017

  22. [22]

    Kriging metamodeling in simulation: A review

    J Kleijnen. Kriging metamodeling in simulation: A review. European journal of operational research, 192(3):707–716, 2009. Kriging approach to investigate stick-slip instability 31

    work page 2009

  23. [23]

    Gaussian processes for regression

    C Williams and C Rasmussen. Gaussian processes for regression. In Ad- vances in neural information processing systems , pages 514–520, 1996

    work page 1996

  24. [24]

    Analysis of support vector regression for approximation of complex engineering analyses

    S Clarke, J Griebsch, and T Simpson. Analysis of support vector regression for approximation of complex engineering analyses. Journal of mechanical design, 127(6):1077–1087, 2005

    work page 2005

  25. [25]

    Universal approximation using radial-basis- function networks

    J Park and I Sandberg. Universal approximation using radial-basis- function networks. Neural computation, 3(2):246–257, 1991

    work page 1991

  26. [26]

    An adaptive sampling strategy for kriging metamodel based on Delaunay triangulation and topsis

    P Jiang, Y Zhang, Q Zhou, X Shao, J Hu, and L Shu. An adaptive sampling strategy for kriging metamodel based on Delaunay triangulation and topsis. Applied Intelligence, pages 1–13, 2017

    work page 2017

  27. [27]

    A statistical approach to some basic mine valuation problems on the witwatersrand

    D Krige. A statistical approach to some basic mine valuation problems on the witwatersrand. Journal of the Southern African Institute of Mining and Metallurgy, 52(6):119–139, 1951

    work page 1951

  28. [28]

    Design and analysis of computer experiments

    J Sacks, W Welch, T Mitchell, and H Wynn. Design and analysis of computer experiments. Statistical science, pages 409–423, 1989

    work page 1989

  29. [29]

    Kriging for interpolation in random simu- lation

    W Van Beers and J Kleijnen. Kriging for interpolation in random simu- lation. Journal of the Operational Research Society , 54(3):255–262, 2003

    work page 2003

  30. [30]

    Review and comparison of dry friction force models

    E Pennestr` ı, V Rossi, P Salvini, and P Valentini. Review and comparison of dry friction force models. Nonlinear dynamics, 83(4):1785–1801, 2016

    work page 2016

  31. [31]

    Friction models and friction compensation

    H Olsson, K ˚Astr¨ om, C De Wit, M G¨ afvert, and P Lischinsky. Friction models and friction compensation. Eur. J. Control , 4(3):176–195, 1998

    work page 1998

  32. [32]

    A survey and com- parison of several friction force models for dynamic analysis of multibody mechanical systems

    F Marques, P Flores, JC Claro, and H Lankarani. A survey and com- parison of several friction force models for dynamic analysis of multibody mechanical systems. Nonlinear Dynamics, 86(3):1407–1443, 2016

    work page 2016

  33. [33]

    Single state elastoplastic friction models

    P Dupont, V Hayward, B Armstrong, and F Altpeter. Single state elastoplastic friction models. IEEE Transactions on automatic control , 47(5):787–792, 2002

    work page 2002

  34. [34]

    Spannungsverteilung in plastischen K¨ orpern

    L Prandtl. Spannungsverteilung in plastischen K¨ orpern. InProceedings of the 1st International Congress on Applied Mechanics , pages 43–54, 1924

    work page 1924

  35. [35]

    Elasto-plastic friction model: contact compliance and stiction

    P Dupont, B Armstrong, and V Hayward. Elasto-plastic friction model: contact compliance and stiction. In American Control Conference, 2000. Proceedings of the 2000, volume 2, pages 1072–1077. IEEE, 2000

    work page 2000

  36. [36]

    A new model for control of systems with friction

    C De Wit, H Olsson, K Astrom, and P Lischinsky. A new model for control of systems with friction. IEEE Transactions on automatic control, 40(3):419–425, 1995

    work page 1995

  37. [37]

    The effect of Coulomb friction and stiction on force control

    W Townsend and Jr Salisbury. The effect of Coulomb friction and stiction on force control. In Proceedings. 1987 IEEE International Conference on Robotics and Automation, volume 4, pages 883–889. IEEE, 1987

    work page 1987

  38. [38]

    Die wesentlichen Eigenschaften der Gleit-und Rollenlager

    R Stribeck. Die wesentlichen Eigenschaften der Gleit-und Rollenlager. Zeitschrift des Vereines Deutscher Ingenieure , 46:1341–1348, 1902

    work page 1902

  39. [39]

    A multiplicative ergodic theorem

    V Oseledec. A multiplicative ergodic theorem. Liapunov characteristic number for dynamical systems. Trans. Moscow Math. Soc. , 19:197–231, 1968

    work page 1968

  40. [40]

    A practical method for cal- culating largest Lyapunov exponents from small data sets

    M Rosenstein, J Collins, and C De Luca. A practical method for cal- culating largest Lyapunov exponents from small data sets. Physica D: Nonlinear Phenomena, 65(1-2):117–134, 1993. 32 Jan N. Fuhg, Am´ elie Fau

    work page 1993

  41. [41]

    A numerical approach to ergodic problem of dissipative dynamical systems

    I Shimada and T Nagashima. A numerical approach to ergodic problem of dissipative dynamical systems. Progress of theoretical physics, 61(6):1605– 1616, 1979

    work page 1979

  42. [42]

    Lyapunov characteris- tic exponents for smooth dynamical systems and for Hamiltonian systems; a method for computing all of them

    G Benettin, L Galgani, A Giorgilli, and J Strelcyn. Lyapunov characteris- tic exponents for smooth dynamical systems and for Hamiltonian systems; a method for computing all of them. part 1: Theory. Meccanica, 15(1):9– 20, 1980

    work page 1980

  43. [43]

    A robust method to estimate the maximal Lyapunov exponent of a time series

    H Kantz. A robust method to estimate the maximal Lyapunov exponent of a time series. Physics letters A , 185(1):77–87, 1994

    work page 1994

  44. [44]

    Quantifying chaos with Lyapunov exponents

    A Wolf. Quantifying chaos with Lyapunov exponents. Chaos, 16:285–317, 1986

    work page 1986

  45. [45]

    Simple chaotic flows with one stable equilibrium

    M Molaie, S Jafari, J Sprott, and S Golpayegani. Simple chaotic flows with one stable equilibrium. International Journal of Bifurcation and Chaos , 23(11):1350188, 2013

    work page 2013

  46. [46]

    Principles of geostatistics

    G Matheron. Principles of geostatistics. Economic geology, 58(8):1246– 1266, 1963

    work page 1963

  47. [47]

    Universal kriging and cokriging as a regression procedure

    A Stein and LCA Corsten. Universal kriging and cokriging as a regression procedure. Biometrics, pages 575–587, 1991

    work page 1991

  48. [48]

    A Bayesian analysis of kriging

    M Handcock and M Stein. A Bayesian analysis of kriging. Technometrics, 35(4):403–410, 1993

    work page 1993

  49. [49]

    Spatial variation: Meddelanden fran statens skogsforskningsin- stitut

    B Mat´ ern. Spatial variation: Meddelanden fran statens skogsforskningsin- stitut. Lecture Notes in Statistics , 36:21, 1960

    work page 1960

  50. [50]

    Metamodel-based importance sampling for structural reliability analysis

    V Dubourg, B Sudret, and F Deheeger. Metamodel-based importance sampling for structural reliability analysis. Probabilistic Engineering Me- chanics, 33:47–57, 2013

    work page 2013

  51. [51]

    Gradient-enhanced kriging for high-dimensional problems

    M Bouhlel and J Martins. Gradient-enhanced kriging for high-dimensional problems. Engineering with Computers , 35(1):157–173, 2019

    work page 2019

  52. [52]

    The development of a hybridized particle swarm for kriging hyperparameter tuning

    D Toal, N Bressloff, A Keane, and C Holden. The development of a hybridized particle swarm for kriging hyperparameter tuning. Engineering optimization, 43(6):675–699, 2011

    work page 2011

  53. [53]

    The design and analysis of computer experiments

    T Santner, B Williams, and W Notz. The design and analysis of computer experiments. Springer Science & Business Media, 2013

    work page 2013

  54. [54]

    Efficient global optimization of expensive black-box functions

    D Jones, M Schonlau, and W Welch. Efficient global optimization of expensive black-box functions. Journal of Global optimization , 13(4):455– 492, 1998

    work page 1998

  55. [55]

    Mul- tidimensional sequential sampling for NURBs-based metamodel develop- ment

    Cameron J Turner, Richard H Crawford, and Matthew I Campbell. Mul- tidimensional sequential sampling for NURBs-based metamodel develop- ment. Engineering with Computers , 23(3):155–174, 2007

    work page 2007

  56. [56]

    A balanced sequential design strategy for global surrogate modeling

    P Singh, D Deschrijver, and T Dhaene. A balanced sequential design strategy for global surrogate modeling. In Simulation Conference (WSC), 2013 Winter, pages 2172–2179. IEEE, 2013

    work page 2013

  57. [57]

    An adaptive sampling approach for kriging metamodeling by maximizing expected prediction error

    H Liu, J Cai, and Y Ong. An adaptive sampling approach for kriging metamodeling by maximizing expected prediction error. Computers & Chemical Engineering, 106:171–182, 2017

    work page 2017

  58. [58]

    Predictive approaches for choosing hy- perparameters in Gaussian processes

    S Sundararajan and S Keerthi. Predictive approaches for choosing hy- perparameters in Gaussian processes. In Advances in neural information Kriging approach to investigate stick-slip instability 33 processing systems, pages 631–637, 2000

    work page 2000

  59. [59]

    Sequential adaptive designs in computer experiments for response surface model fit

    C Lam. Sequential adaptive designs in computer experiments for response surface model fit. PhD thesis, The Ohio State University, 2008

    work page 2008

  60. [60]

    An innovative adaptive kriging approach for effi- cient binary classification of mechanical problems

    JN Fuhg and A Fau. An innovative adaptive kriging approach for effi- cient binary classification of mechanical problems. manuscript submitted to publication, in 2019

    work page 2019

  61. [61]

    Adaptive surrogate models for parametric studies

    JN Fuhg. Adaptive surrogate models for parametric studies. Master’s thesis, Leibniz Universit¨ at Hannover, Arxiv platform https://arxiv.org/abs/1905.05345, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1905

  62. [62]

    An algorithm for fast optimal latin hypercube design of experiments

    F Viana, G Venter, and V Balabanov. An algorithm for fast optimal latin hypercube design of experiments. International journal for numerical methods in engineering, 82(2):135–156, 2010

    work page 2010