Pith. sign in

REVIEW 3 major objections 4 minor 73 references

A 21-parameter closed-form material law lets designers optimize both topology and composition of 3D-printed parts in a single end-to-end loop.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 05:57 UTC pith:ZLF6NH2N

load-bearing objection A competent, reproducible integration of a sparsified PANN constitutive law into FEniCSx topology optimization, with a real practical contribution to finite-strain void handling—but the design claims rest on a borrowed, under-validated material model. the 3 major comments →

arxiv 2607.13174 v1 pith:ZLF6NH2N submitted 2026-07-14 physics.comp-ph cs.CEcs.RO

Towards end-to-end optimization in multimaterial 3D printing

Xue-Ling Luo , Steven Yang , Jingye Tan , Robert F. Shepherd , Noy Cohen , Nikolaos Bouklas This is my paper
classification physics.comp-ph cs.CEcs.RO MSC 74P1574B2068T0774S05
keywords multimaterial 3D printingtopology optimizationfunctionally graded materialsphysics-augmented neural networkshyperelasticityadjoint state methodsoft roboticsdigital materials
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper's thesis is that multimaterial 3D printing need not rely on trial-and-error prototyping: a compact, composition-aware hyperelastic law—extracted from scarce experimental data by a sparsified physics-augmented neural network and written out as a 21-parameter explicit expression—can be inserted directly into a finite-element optimization loop because the solver differentiates the expression symbolically rather than through a network. This removes the main bottleneck that has kept neural-network constitutive models out of topology optimization. The framework is demonstrated on soft robotic gripper components, where continuous material blending is tuned for anisotropic contact response, and, in the central result, topology and material composition are optimized concurrently under a non-failure stretch constraint. The concurrent design achieves the best global stiffness among the constrained scenarios. If correct, the pipeline replaces manual constitutive-model fitting with data-driven, end-to-end design of functionally graded parts.

Core claim

The authors establish that a sparsified physics-augmented neural network (pICNN) can be reduced to a closed-form strain-energy function of the isochoric invariants and the local composition ratio c, with just 21 nonzero parameters, so that stresses, tangent operators, objective functions, and adjoint sensitivities are all obtained by exact symbolic differentiation inside a finite element framework. Using this law within a near-incompressible penalty formulation (bulk modulus from a fitted composition-dependent tensile modulus, Poisson ratio 0.49), they run adjoint-based topology and material-distribution optimization. They demonstrate that continuous composition fields alone can create stron

What carries the argument

The enabling object is the 21-parameter closed-form pICNN strain-energy density (Eqs. 2–3), a composition-aware hyperelastic law of the isochoric invariants Ī1, Ī2 and the mix ratio c. Because the expression uses only exponentials, logarithms, and powers, the finite element framework can symbolically differentiate it to form explicit stress and tangent tensors; this is what allows the adjoint state method to compute exact gradients with respect to the many nodal design variables without automatic differentiation through a network. A second supporting mechanism is the void-interpolation scheme (Eqs. 14–17) that keeps the near-incompressible solver stable in void regions and is explicitly deco

Load-bearing premise

The load-bearing premise is that the closed-form 21-parameter hyperelastic law taken from prior work describes the real digital material accurately for every composition and deformation state the optimizer will explore, a premise the paper does not re-validate and that rests partly on limited experimental data.

What would settle it

3D-print the optimized gripper finger (or its topology and composition fields) and subject it to the finite clamp deformation; if the measured global stiffness or the stretch at failure deviates sharply from the numerical prediction—specifically, if the I1-based non-failure constraint is violated below the simulated threshold—the end-to-end claim fails. A cheaper numerical falsifier is to re-run the optimized design with the original full pICNN model or with an independent hyperelastic fit and check whether stiffness and constraint satisfaction match the sparse-law results.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Multimaterial parts can be designed from scarce material-test data—on the order of tens of experiments—without hand-fitting a constitutive model inside the optimizer.
  • Continuous composition fields become first-class design variables for topology optimization, enabling functionally graded structures that were previously left to intuition.
  • Exact symbolic differentiation of the constitutive law removes the need for Jacobian-free solvers and their convergence problems, making nonlinear finite element optimization faster and more robust for learned material models.
  • Concurrent topology and composition optimization can satisfy non-failure stretch constraints with less stiffness sacrifice than optimizing either field alone.
  • The formulation is dimension-independent and directly applicable to 3D problems without modification.

Where Pith is reading between the lines

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

  • A testable extension is to replace the fixed I1-based failure criterion with composition- and deformation-dependent damage data; the framework's explicit derivatives would accommodate any smooth replacement.
  • The optimization may be sensitive to the accuracy of the pICNN law at extreme compositions or deformations: validating the optimized designs with the original unsparsified network or with new experiments would bound the error introduced by the 21-parameter truncation.
  • Because the workflow decouples constitutive modeling from optimization, the same pipeline could be applied to other graded material systems (e.g., different base polymers or embedded fillers) by re-fitting the sparse law once.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper presents an end-to-end computational framework for multimaterial 3D printing that couples a sparsified physics-augmented neural network (pICNN) hyperelastic constitutive law—taken from the authors' prior work [36]—with topology and continuous composition optimization. The closed-form expression of the strain energy density enables exact symbolic differentiation within FEniCSx, and the adjoint state method is used to compute sensitivities with respect to density and composition fields. After detailing the numerical ingredients (nearly incompressible penalty formulation, PDE-based filtering, third medium contact, and MMA updates), the paper demonstrates the workflow on (i) a trapezoidal contact point under small deformation, (ii) a semicircular contact point under finite-deformation inclined compression, and (iii) a soft gripper finger under concurrent topology/composition optimization with an I1-based failure constraint. The central claim is that the explicit, 21-parameter constitutive law removes the differentiation bottleneck of ML constitutive models and that concurrent optimization outperforms composition-only or topology-only optimization in the gripper finger example.

Significance. If the constitutive law is reliable in the regimes visited by the optimizer, this is a useful step toward practical design with data-driven hyperelastic models: the closed-form energy enables exact symbolic adjoint sensitivities, the PDE filtering is standard, and the released code makes the numerical machinery reproducible. The internal comparison across four scenarios in §3.3 is a clear and appropriate way to isolate the benefit of concurrent optimization, and the reported superiority of the concurrent design is plausible. The paper also correctly identifies the bottleneck of differentiating neural-network constitutive models and offers a concrete remedy. However, the significance of the quantitative results is contingent on the accuracy and range of validity of the borrowed material law and failure criterion, which are not revalidated here.

major comments (3)
  1. [§2.1.2, Eqs. (2)-(6), §3.3, Eq. (29)] The entire demonstration inherits the pICNN law of Eqs. (2)-(3) and the fitted relations E(c) and I1^crit(c) from [36], with no revalidation against independent experiments or hold-out data. The manuscript itself concedes in §2.1.2 that volumetric data were inconclusive, yet the nearly incompressible penalty uses ν=0.49 with no direct volumetric measurement. Moreover, the optimized gripper finger undergoes non-uniform multiaxial and compressive finite deformation, whereas [36] was trained on uniaxial tension and torsion. If the law extrapolates poorly in the composition or stretch ranges preferred by the optimizer, every optimized design in §3 is affected, including the claim of 'superior' concurrent stiffness in Fig. 3(b). At minimum, add a validation study (e.g., multiaxial experiments or comparison against an independent data set) or explicitly reposition the examples as numerical pro
  2. [§3.3, Eqs. (29)-(31)] The 'non-failure' constraint is not strictly enforced locally. I1^crit(c) is fitted to only five experimental stretch limits (Fig. B.5), the safety factor n_safe=10 is heuristic, and the global functional constraint allows a normalized violation integral of 0.25. As the text acknowledges, this targets only a 75% reduction of the high-risk area, so local regions above the critical invariant can remain. The phrase 'non-failure stretch constraints' overstates what is implemented. Please report sensitivity of the optimized designs to n_safe and the 0.25 tolerance, and clarify the actual level of local constraint satisfaction in the reported designs.
  3. [§3.2 and Appendix A] The contact optimization results depend on the third medium parameters k_3rd=10^-5 and k_reg=10^-3, which appear without a sensitivity study. Since the anisotropic reaction-force ratios in Fig. 2(c) are the main quantitative output of this section, a parameter sweep or an alternative contact formulation would help establish that the optimized material distribution—rather than the contact regularization—is responsible for the observed behavior. This is especially important because the third medium adds a nonstandard regularization term (Eq. A.3/A.4) whose effect on the optimization landscape has not been assessed.
minor comments (4)
  1. [§2.3.2, Eq. (17)] The notation in Eq. (17) is inconsistent with Eq. (14): Eq. (14) uses ψ(Iu) and ψ_L(Iu), while Eq. (17) writes ψ(I + I ∂u/∂X, c). Please clarify the argument and ensure the displacement-gradient scaling is reported consistently.
  2. [Algorithm 1, line 13] The convergence criterion 'decreases by less than TOL in two consecutive iterations or has not decreased for 50 iterations after 500 iterations' is ambiguous. Does the 500-iteration condition apply only after iteration 500, or is it a separate stopping rule? Please rewrite for clarity.
  3. [Fig. 1(b)] The vertical axis is 'reaction forces normalized with respect to the initial homogeneous design', but it is not immediately clear whether the plotted values are horizontal and vertical forces separately or a combined metric. Please label the curves directly (e.g., |F_x| and |F_y|) to avoid ambiguity.
  4. [References] Reference [40] is a preprint; provide a versioned DOI or archival identifier so the exact FEniCSx version can be reproduced.

Circularity Check

0 steps flagged

No circular derivation: constitutive law and failure data enter as external inputs from [36]; optimized designs are genuine outputs of the optimization, not by-construction equivalents of the fitted relations.

full rationale

The claimed derivation chain is not circular. The pICNN strain energy (Eqs. (2)-(3)) is imported verbatim from the authors' prior work [36], which calibrates it against 30 uniaxial tension and torsion experimental datasets. This is a data-derived external input to the present optimization framework, not a quantity defined by the paper's own outputs. The fitted tensile-modulus relation E(c) (Eq. (5)) is obtained by symbolically differentiating Eq. (2) in the infinitesimal regime and is used only to set the penalty bulk modulus in Eq. (6) for the near-incompressible penalty; this is an internal consistency calibration of the penalty parameter to the base law, not a prediction that is then validated against the same fit. Likewise, the failure criterion I1^crit(c) (Eq. (29)) is a least-squares fit to five experimental stretch limits reported in [36] and is used as a constraint input; the fact that the constrained optimizations in Section 3.3 satisfy g2 is by construction of constrained optimization, but the paper's central comparative claim (concurrent ρ-c optimization achieves the highest stiffness among the four scenarios, Fig. 3b) is an output of the optimizer under identical inputs, not imposed by the fit. The adjoint state method, Helmholtz filtering, MMA updates, and FEniCSx symbolic differentiation are all derived self-contained within Sections 2.2-2.5. The paper's own concession in Section 2.1.2 that 'volumetric data were inconclusive' flags a model-validity and extrapolation risk, not a circular step: if the borrowed law fails in multiaxial finite deformation or at extrapolated compositions, every optimized design inherits that error, but the error is not an equivalence between inputs and outputs. No step reduces to its own input by definition.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 1 invented entities

The paper's own new contribution (the optimizer) imports the material law and the failure data from previous work, and then adds hand-chosen regularization parameters (bulk modulus proxy, n_safe, 0.25 tolerance, third medium constants). The central demonstrates therefore rest on the fitted/borrowed inputs. No new physical entity is proposed; the 'third medium' is a standard numerical device.

free parameters (4)
  • Bulk modulus penalty parameter κ via fitted tensile modulus E(c) = E(c) = 0.304c^4 - 0.962c^3 + 0.899c^2 + 0.399c + 0.697 (Eq. 5); κ = E(c)/(3(1-2ν)) with ν=0.49
    Least-squares fit to symbolically-derived infinitesimal tensile modulus; converts the exact-incompressibility model into a near-incompressible penalty model used in all optimizations.
  • Critical invariant relation I1^crit(c) = I1^crit(c) = 0.224c^2 - 1.125c + 4.884 (Eq. 29)
    Least-squares fit to five experimental uniaxial stretch limits from [36]; defines the safety constraint in §3.3.
  • Penalty/safety factors and tolerances = n_safe=10, constraint tolerance 0.25, ρ_min=0.001, ϱ∈[2,5], θ_thres=0.01, θ_sharp=500
    Chosen by hand/convention to stabilize the optimization and relax the failure constraint; they directly shape the reported objective/constraint tradeoff.
  • Third medium parameters for contact examples = k_3rd=1e-5, k_reg=1e-3 (in strain-energy units)
    Chosen to obtain convergence in the finite-deformation contact example; sensitivity of the optimized composition field to these values is not reported.
axioms (4)
  • domain assumption The closed-form pICNN strain energy of Eq. (2)-(3), with its 21 parameters, accurately represents the hyperelastic response of Agilus/Digital-ABS blends for all c∈[0,1] and all deformation states encountered in optimization.
    Fitted to a limited set of uniaxial/torsion experiments in prior work [36]; the present paper never revalidates it and even notes volumetric data were inconclusive (§2.1.2).
  • domain assumption Infinitesimal-strain-based E(c) fit can be extended via isotropic elasticity with ν=0.49 to a bulk modulus for the penalty model at finite strains.
    Eqs. (5)-(6) in §2.1.2; the incompressible formulation is replaced by a near-incompressible penalty term without a finite-strain calibration.
  • domain assumption An I1-based eight-chain criterion calibrated from uniaxial stretch limits is a valid local failure criterion under multiaxial loading with a safety factor of 10.
    §3.3, Eqs. (29)-(31); the five λ^crit values are from [36] and converted to I1^crit assuming uniaxial kinematics.
  • domain assumption Plane-strain 2D simulations of quasi-static hyperelastic response are representative of the 3D soft robot gripper behavior of interest.
    Throughout §3 and explicitly in §2.5 ('all numerical examples... in 2D'), with no experimental validation.
invented entities (1)
  • Third medium contact domain with regularized gradient energy no independent evidence
    purpose: Numerical device to model finite-deformation contact without explicit contact mechanics; used in the semicircular fingertip example (§3.2, Appendix A).
    A computational regularization taken from refs. [69, 70]; it introduces fictitious knobs (k_3rd, k_reg) and a regularization energy that have no physical meaning outside the discretization.

pith-pipeline@v1.3.0-alltime-deepseek · 20709 in / 8312 out tokens · 63286 ms · 2026-08-02T05:57:51.141885+00:00 · methodology

0 comments
read the original abstract

Multimaterial 3D printing enables the fabrication of functionally graded components, but optimizing their spatial material distribution alongside structural topology remains a formidable challenge due to high-dimensional design spaces and complex constitutive modeling. This paper presents an end-to-end computational framework integrating sparsified physics-augmented neural networks with finite-element-based topology optimization. By extracting closed-form, composition-aware hyperelastic constitutive laws from experimental data, this approach facilitates exact symbolic differentiation via the adjoint state method implemented with FEniCSx, efficiently circumventing the bottlenecks of applying neural network constitutive models. This pipeline is deployed on soft robotic gripper applications, demonstrating continuous composition optimization for highly anisotropic contact responses, and the concurrent optimization of macroscopic topology and material distribution under non-failure stretch constraints. This methodology could replace laborious empirical prototyping, establishing interpretable machine-learning models as practical, robust design primitives for advanced multimaterial additive manufacturing.

Figures

Figures reproduced from arXiv: 2607.13174 by Jingye Tan, Nikolaos Bouklas, Noy Cohen, Robert F. Shepherd, Steven Yang, Xue-Ling Luo.

Figure 1
Figure 1. Figure 1: Optimization of the continuous composition ratio [PITH_FULL_IMAGE:figures/full_fig_p022_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Optimization of the continuous composition ratio [PITH_FULL_IMAGE:figures/full_fig_p025_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Optimization of the continuous composition ratio and structural topology for a soft robotic finger [PITH_FULL_IMAGE:figures/full_fig_p030_3.png] view at source ↗

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

73 extracted references · 4 canonical work pages

  1. [1]

    S.-J. Ahn, H. Lee, K.-J. Cho, 3D printing with a 3D printed digital material filament for programming functional gradients, Nature Communications 15 (1) (2024) 3605.doi:10.1038/s41467-024-47480-5

    work page doi:10.1038/s41467-024-47480-5 2024

  2. [2]

    Rossetti, L

    L. Rossetti, L. Kuntz, E. Kunold, J. Schock, K. M¨ uller, H. Grabmayr, J. Stolberg-Stolberg, F. Pfeiffer, S. Sieber, R. Burgkart, et al., The microstructure and micromechanics of the tendon–bone insertion, Nature materials 16 (6) (2017) 664–670

    2017

  3. [3]

    Pragya, T

    A. Pragya, T. K. Ghosh, Soft Functionally Gradient Materials and Structures – Natural and Manmade: A Review, Advanced Materials 35 (49) (2023) 2300912.doi:10.1002/adma.202300912

    work page doi:10.1002/adma.202300912 2023

  4. [4]

    N. W. Bartlett, M. T. Tolley, J. T. Overvelde, J. C. Weaver, B. Mosadegh, K. Bertoldi, G. M. Whitesides, R. J. Wood, A 3d-printed, functionally graded soft robot powered by combustion, Sci- ence 349 (6244) (2015) 161–165

    2015

  5. [5]

    Kumar, K

    S. Kumar, K. M. Reddy, A. Kumar, G. R. Devi, Development and characterization of polymer–ceramic continuous fiber reinforced functionally graded composites for aerospace application, Aerospace Science and Technology 26 (1) (2013) 185–191

    2013

  6. [6]

    Saleh, J

    B. Saleh, J. Jiang, R. Fathi, T. Al-Hababi, Q. Xu, L. Wang, D. Song, A. Ma, 30 years of function- ally graded materials: An overview of manufacturing methods, applications and future challenges, Composites Part B: Engineering 201 (2020) 108376

    2020

  7. [7]

    K. Li, M. Zhang, Z. Zhang, P. Jin, Y. Wang, W. Yan, L. Zhu, D. Z. Zhang, L. E. Murr, High per- formance realization of functionally graded materials based on integrated optimal design and additive manufacturing: A review, International Materials Reviews 70 (6) (2025) 497–547

    2025

  8. [8]

    Muthuram, P

    N. Muthuram, P. S. Madhav, D. K. Vasan, M. E. Mohan, G. Prajeeth, A review of recent literatures in poly jet printing process, Materials Today: Proceedings 68 (2022) 1906–1920

    2022

  9. [9]

    X. Wei, A. Bhardwaj, C.-C. Shih, L. Zeng, B. Tai, Z. Pei, Experimental investigation of stratasys j750 polyjet printer: Effects of orientation and layer thickness on mechanical properties, in: Interna- tional Manufacturing Science and Engineering Conference, Vol. 58745, American Society of Mechanical Engineers, 2019, p. V001T01A021

    2019

  10. [10]

    C. Wade, D. Beck, R. MacCurdy, Implicit modeling for 3d-printed multi-material computational object design via python, in: Proceedings of the ACM Symposium on Computational Fabrication, 2025, pp. 1–16

    2025

  11. [11]

    Smith, T

    L. Smith, T. Hainsworth, Z. Jordan, X. Bell, R. MacCurdy, A seamless workflow for design and fab- rication of multimaterial pneumatic soft actuators, in: 2021 IEEE 17th International Conference on Automation Science and Engineering (CASE), IEEE, 2021, pp. 718–723

    2021

  12. [12]

    Nazir, O

    A. Nazir, O. Gokcekaya, K. M. M. Billah, O. Ertugrul, J. Jiang, J. Sun, S. Hussain, Multi-material additive manufacturing: A systematic review of design, properties, applications, challenges, and 3d 36 printing of materials and cellular metamaterials, Materials & Design 226 (2023) 111661

    2023

  13. [13]

    Hirano, Modeling and design of micro-structures: focusing on functionally graded materials and future prospects, Frontiers in Materials 12 (2025) 1659727

    T. Hirano, Modeling and design of micro-structures: focusing on functionally graded materials and future prospects, Frontiers in Materials 12 (2025) 1659727

    2025

  14. [14]

    A. Garg, W. Zheng, R. Raman, L. Li, Machine learning in functionally graded materials and nano fgms: A comprehensive review of predictive modeling for mechanical behavior: A. garg et al., Archives of Computational Methods in Engineering 33 (1) (2026) 533–575

    2026

  15. [15]

    Bocciarelli, G

    M. Bocciarelli, G. Bolzon, G. Maier, A constitutive model of metal–ceramic functionally graded material behavior: formulation and parameter identification, Computational Materials Science 43 (1) (2008) 16– 26

    2008

  16. [16]

    J. P. Pascon, Large deformation analysis of functionally graded visco-hyperelastic materials, Computers & Structures 206 (2018) 90–108

    2018

  17. [17]

    L. Xue, G. Dui, B. Liu, L. Xin, A phenomenological constitutive model for functionally graded porous shape memory alloy, International journal of engineering science 78 (2014) 103–113

    2014

  18. [18]

    Hornik, M

    K. Hornik, M. Stinchcombe, H. White, Multilayer feedforward networks are universal approximators, Neural networks 2 (5) (1989) 359–366

    1989

  19. [19]

    L. Lu, P. Jin, G. E. Karniadakis, Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators, arXiv preprint arXiv:1910.03193 (2019)

    Pith/arXiv arXiv 1910

  20. [20]

    J. N. Fuhg, G. Anantha Padmanabha, N. Bouklas, B. Bahmani, W. Sun, N. N. Vlassis, M. Flaschel, P. Carrara, L. De Lorenzis, A Review on Data-Driven Constitutive Laws for Solids, Archives of Com- putational Methods in Engineering 32 (3) (2025) 1841–1883.doi:10.1007/s11831-024-10196-2

    doi:10.1007/s11831-024-10196-2 2025

  21. [21]

    Kirchdoerfer, M

    T. Kirchdoerfer, M. Ortiz, Data-driven computational mechanics, Computer Methods in Applied Me- chanics and Engineering 304 (2016) 81–101.doi:10.1016/j.cma.2016.02.001

    doi:10.1016/j.cma.2016.02.001 2016

  22. [22]

    S. Tang, G. Zhang, H. Yang, Y. Li, W. K. Liu, X. Guo, MAP123: A data-driven approach to use 1D data for 3D nonlinear elastic materials modeling, Computer Methods in Applied Mechanics and Engineering 357 (2019) 112587.doi:10.1016/j.cma.2019.112587

    arXiv 2019

  23. [23]

    Luo, J.-Y

    X.-L. Luo, J.-Y. Ye, P.-S. Ma, L.-W. Zhang, Data-driven enhanced phase field models for highly accurate prediction of mode i and mode ii fracture, Computer Methods in Applied Mechanics and Engineering 400 (2022) 115535

    2022

  24. [24]

    Guo, A data-driven approach to predicting the anisotropic mechanical behaviour of voided single crystals, Journal of the Mechanics and Physics of Solids (2022)

    H.-J. Guo, A data-driven approach to predicting the anisotropic mechanical behaviour of voided single crystals, Journal of the Mechanics and Physics of Solids (2022)

    2022

  25. [25]

    D. K. Klein, M. Fern´ andez, R. J. Martin, P. Neff, O. Weeger, Polyconvex anisotropic hyperelasticity with neural networks, Journal of the Mechanics and Physics of Solids 159 (2022) 104703

    2022

  26. [26]

    K. Xu, D. Z. Huang, E. Darve, Learning constitutive relations using symmetric positive definite neural 37 networks, Journal of Computational Physics 428 (2021) 110072.doi:10.1016/j.jcp.2020.110072

    arXiv 2021

  27. [27]

    Linka, M

    K. Linka, M. Hillg¨ artner, K. P. Abdolazizi, R. C. Aydin, M. Itskov, C. J. Cyron, Constitutive artificial neural networks: A fast and general approach to predictive data-driven constitutive modeling by deep learning, Journal of Computational Physics 429 (2021) 110010

    2021

  28. [28]

    Ciftci, K

    K. Ciftci, K. Hackl, A physics-informed GAN framework based on model-free data-driven computational mechanics, Computer Methods in Applied Mechanics and Engineering 424 (2024) 116907.doi:10. 1016/j.cma.2024.116907

    arXiv 2024

  29. [29]

    J. N. Fuhg, N. Bouklas, On physics-informed data-driven isotropic and anisotropic constitutive mod- els through probabilistic machine learning and space-filling sampling, Computer Methods in Applied Mechanics and Engineering 394 (2022) 114915

    2022

  30. [30]

    J. N. Fuhg, N. Bouklas, R. E. Jones, Stress representations for tensor basis neural networks: alternative formulations to finger–rivlin–ericksen, Journal of Computing and Information Science in Engineering 24 (11) (2024) 111007

    2024

  31. [31]

    V. Tac, F. S. Costabal, A. B. Tepole, Data-driven tissue mechanics with polyconvex neural ordinary differential equations, Computer Methods in Applied Mechanics and Engineering 398 (2022) 115248

    2022

  32. [32]

    Luo, C.-C

    X.-L. Luo, C.-C. Lyu, L.-W. Zhang, Physics-informed ensemble learning for robustly extrapolating and revealing fatigue life of composites, Composites Science and Technology (2025) 111302

    2025

  33. [33]

    J. A. Janssen, G. Haikal, E. C. DeCarlo, M. J. Hartnett, M. L. Kirby, A physics-informed general convolutional network for the computational modeling of materials with damage, Journal of Computing and Information Science in Engineering 24 (11) (2024) 111002

    2024

  34. [34]

    X. Luo, J. Huang, L. Shen, H. Wang, Z. Shi, Physics-informed neural networks for constitutive model- ing and multiphysics coupling in viscoelastic materials: Applications to asphalt pavement mechanics, Neural Networks (2026) 108803

    2026

  35. [35]

    Garbrecht, M

    K. Garbrecht, M. Aguilo, A. Sanderson, A. Rollett, R. M. Kirby, J. Hochhalter, Interpretable Machine Learning for Texture-Dependent Constitutive Models with Automatic Code Generation for Topological Optimization, Integrating Materials and Manufacturing Innovation 10 (3) (2021) 373–392.doi:10. 1007/s40192-021-00231-6

    2021

  36. [36]

    S. Yang, M. Levin, G. Anantha Padmanabha, M. Borshevsky, O. Cohen, D. T. Seidl, R. E. Jones, N. Bouklas, N. Cohen, Physics augmented machine learning discovery of composition-dependent con- stitutive laws for 3D printed digital materials, International Journal of Engineering Science 217 (2025) 104381.doi:10.1016/j.ijengsci.2025.104381

    arXiv 2025

  37. [37]

    Revels, M

    J. Revels, M. Lubin, T. Papamarkou, Forward-mode automatic differentiation in julia, arXiv preprint arXiv:1607.07892 (2016)

    Pith/arXiv arXiv 2016

  38. [38]

    D. A. Knoll, D. E. Keyes, Jacobian-free newton–krylov methods: a survey of approaches and applica- 38 tions, Journal of Computational Physics 193 (2) (2004) 357–397

    2004

  39. [39]

    S. L. Brunton, J. L. Proctor, J. N. Kutz, Discovering governing equations from data by sparse identifi- cation of nonlinear dynamical systems, Proceedings of the national academy of sciences 113 (15) (2016) 3932–3937

    2016

  40. [40]

    I. A. Baratta, J. P. Dean, J. S. Dokken, M. Habera, J. S. Hale, C. N. Richardson, M. E. Rognes, M. W. Scroggs, N. Sime, G. N. Wells, DOLFINx: the next generation FEniCS problem solving environment, preprint (2023).doi:10.5281/zenodo.10447666

    doi:10.5281/zenodo.10447666 2023

  41. [41]

    Flaschel, S

    M. Flaschel, S. Kumar, L. De Lorenzis, Unsupervised discovery of interpretable hyperelastic constitutive laws, Computer Methods in Applied Mechanics and Engineering 381 (2021) 113852

    2021

  42. [42]

    J. N. Fuhg, R. E. Jones, N. Bouklas, Extreme sparsification of physics-augmented neural networks for interpretable model discovery in mechanics, Computer Methods in Applied Mechanics and Engineering 426 (2024) 116973

    2024

  43. [43]

    Sigmund, K

    O. Sigmund, K. Maute, Topology optimization approaches: A comparative review, Structural and multidisciplinary optimization 48 (6) (2013) 1031–1055

    2013

  44. [44]

    X. Yu, J. Griffis, G. Manogharan, A. Panesar, Multi-material additive manufacturing: a computational design perspective, Virtual and Physical Prototyping 20 (1) (2025) e2546671

    2025

  45. [45]

    Z. Li, H. Xu, S. Zhang, J. Cui, X. Liu, Multi-material structures topology optimization for thin-walled tube used by vehicles under static load: A review: Z. li et al., Archives of Computational Methods in Engineering 32 (8) (2025) 4769–4809

    2025

  46. [46]

    X. S. Zhang, G. H. Paulino, A. S. Ramos Jr, Multi-material topology optimization with multiple volume constraints: a general approach applied to ground structures with material nonlinearity, Structural and Multidisciplinary Optimization 57 (1) (2018) 161–182

    2018

  47. [47]

    F. M. Conde, P. G. Coelho, J. M. Guedes, Multi-scale topology optimization of structures with multi- material microstructures using stiffness and mass design criteria, Advances in Engineering Software 187 (2024) 103566

    2024

  48. [48]

    Y. Guo, T. Gang, S. Cheng, Y. Wang, L. Chen, Multi-scale multi-material concurrent topology opti- mization of graded cellular structures, Engineering with Computers 41 (4) (2025) 2737–2764

    2025

  49. [49]

    Shimoda, J

    M. Shimoda, J. Fujita, M. Al Ali, A. Kamiya, Concurrent multiscale and multi-material optimization method for natural vibration design of porous structures, International Journal for Numerical Methods in Engineering 125 (7) (2024) e7424

    2024

  50. [50]

    J. Wu, O. Sigmund, J. P. Groen, Topology optimization of multi-scale structures: a review, Structural and Multidisciplinary Optimization 63 (3) (2021) 1455–1480

    2021

  51. [51]

    Ma, X.-C

    P.-S. Ma, X.-C. Liu, X.-L. Luo, S. Li, L.-W. Zhang, Asymptotic homogenization of phase-field fracture model: An efficient multiscale finite element framework for anisotropic fracture, International Journal 39 for Numerical Methods in Engineering 125 (13) (2024) e7489

    2024

  52. [52]

    I. F. Ituarte, N. Boddeti, V. Hassani, M. L. Dunn, D. W. Rosen, Design and additive manufacture of functionally graded structures based on digital materials, Additive Manufacturing 30 (2019) 100839. doi:10.1016/j.addma.2019.100839

    arXiv 2019

  53. [53]

    Vijayakumaran, J

    H. Vijayakumaran, J. B. Russ, G. H. Paulino, M. A. Bessa, Consistent machine learning for topology optimization with microstructure-dependent neural network material models, Journal of the Mechanics and Physics of Solids 196 (2025) 106015.doi:10.1016/j.jmps.2024.106015

    arXiv 2025

  54. [54]

    M. Seo, S. Min, Novel material representation method via a deep learning model for multi-scale topology optimization, Advances in Engineering Software 174 (2022) 103300

    2022

  55. [55]

    D. Wei, F. Zeng, J. Cui, Y. Wang, Machine learning-based multiscale topology optimization framework for nonlinear materials, International Journal of Mechanical Sciences 314 (2026) 111380.doi:10.1016/ j.ijmecsci.2026.111380

    arXiv 2026

  56. [56]

    J. Tan, G. A. Padmanabha, S. J. Yang, N. Bouklas, Towards rapid constitutive model discovery from multi-modal data: Physics augmented finite element model updating (pafemu), arXiv preprint arXiv:2604.07746 (2026)

    Pith/arXiv arXiv 2026

  57. [57]

    Barzilai, J

    J. Barzilai, J. M. Borwein, Two-Point Step Size Gradient Methods, IMA Journal of Numerical Analysis 8 (1) (1988) 141–148.doi:10.1093/imanum/8.1.141

    doi:10.1093/imanum/8.1.141 1988

  58. [58]

    Holmberg, B

    E. Holmberg, B. Torstenfelt, A. Klarbring, Stress constrained topology optimization, Structural and Multidisciplinary Optimization 48 (1) (2013) 33–47.doi:10.1007/s00158-012-0880-7

    doi:10.1007/s00158-012-0880-7 2013

  59. [59]

    K. Svanberg, The method of moving asymptotes—a new method for structural optimization, Inter- national Journal for Numerical Methods in Engineering 24 (2) (1987) 359–373.doi:10.1002/nme. 1620240207

    doi:10.1002/nme 1987

  60. [60]

    Svanberg, A class of globally convergent optimization methods based on conservative convex sepa- rable approximations, SIAM journal on optimization 12 (2) (2002) 555–573

    K. Svanberg, A class of globally convergent optimization methods based on conservative convex sepa- rable approximations, SIAM journal on optimization 12 (2) (2002) 555–573

    2002

  61. [61]

    S. G. Johnson, The NLopt nonlinear-optimization package,https://github.com/stevengj/nlopt (2007)

    2007

  62. [62]

    M. P. Bendsøe, Optimal shape design as a material distribution problem, Structural optimization 1 (4) (1989) 193–202

    1989

  63. [63]

    Sigmund, A 99 line topology optimization code written in matlab, Structural and multidisciplinary optimization 21 (2) (2001) 120–127

    O. Sigmund, A 99 line topology optimization code written in matlab, Structural and multidisciplinary optimization 21 (2) (2001) 120–127

    2001

  64. [64]

    F. Wang, B. S. Lazarov, O. Sigmund, J. S. Jensen, Interpolation scheme for fictitious domain techniques and topology optimization of finite strain elastic problems, Computer Methods in Applied Mechanics and Engineering 276 (2014) 453–472.doi:10.1016/j.cma.2014.03.021

    work page doi:10.1016/j.cma.2014.03.021 2014

  65. [65]

    S. Deng, H. D. Espinosa, W. Chen, Data-driven multiscale nonlinear topology optimization of func- 40 tionally graded soft composites under large deformations, Computational Mechanics (Dec. 2025). doi:10.1007/s00466-025-02730-1

    work page doi:10.1007/s00466-025-02730-1 2025

  66. [66]

    B. S. Lazarov, O. Sigmund, Filters in topology optimization based on Helmholtz-type differen- tial equations, International Journal for Numerical Methods in Engineering 86 (6) (2011) 765–781. doi:10.1002/nme.3072

    doi:10.1002/nme.3072 2011

  67. [67]

    F. L. Hammond, J. Weisz, A. A. de la Llera Kurth, P. K. Allen, R. D. Howe, Towards a design optimization method for reducing the mechanical complexity of underactuated robotic hands, in: 2012 IEEE International conference on robotics and automation, IEEE, 2012, pp. 2843–2850

    2012

  68. [68]

    N. D. Mankame, G. Ananthasuresh, Contact aided compliant mechanisms: concept and preliminaries, in: International design engineering technical conferences and computers and information in engineering conference, Vol. 36533, American Society of Mechanical Engineers, 2002, pp. 109–121

    2002

  69. [69]

    Faltus, M

    O. Faltus, M. Hor´ ak, M. Doˇ sk´ aˇ r, O. Rokoˇ s, Third medium finite element contact formulation for pneumatically actuated systems, Computer Methods in Applied Mechanics and Engineering 431 (2024) 117262.doi:10.1016/j.cma.2024.117262

    arXiv 2024

  70. [70]

    Wriggers, J

    P. Wriggers, J. Korelc, Ph. Junker, A third medium approach for contact using first and second order finite elements, Computer Methods in Applied Mechanics and Engineering 436 (2025) 117740.doi: 10.1016/j.cma.2025.117740

    arXiv 2025

  71. [71]

    E. M. Arruda, M. C. Boyce, A three-dimensional constitutive model for the large stretch behavior of rubber elastic materials, Journal of the Mechanics and Physics of Solids 41 (2) (1993) 389–412

    1993

  72. [72]

    S. M. Mousavi, J. Mulderrig, B. Talamini, N. Bouklas, A chain stretch-based gradient-enhanced model for damage and fracture in elastomers, Computer Methods in Applied Mechanics and Engineering 444 (2025) 118103

    2025

  73. [73]

    G. L. Bluhm, O. Sigmund, K. Poulios, Internal contact modeling for finite strain topology optimization, Computational Mechanics 67 (4) (2021) 1099–1114.doi:10.1007/s00466-021-01974-x. 41

    doi:10.1007/s00466-021-01974-x 2021