A nonlocal model for heterogeneous material flow on conveyor belts

Fabian Ziegler; Paola Goatin; Simone G\"ottlich

REVIEW 4 major objections 5 minor 1 cited by

A nonlocal model for heterogeneous material flow on conveyor belts

T0 review · 4 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read For a two-dimensional nonlocal system describing conveyor-belt flow of mixed particle classes, this paper proves existence, uniqueness, and Lipschitz stability of entropy solutions via a convergent Roe finite-volume scheme.

desk verdict Careful extension of the scalar Roe scheme to nonlocal systems with obstacles, but the headline well-posedness claim covers only a smoothed Heaviside, and a key entropy estimate is asserted without proof. read the letter →

arxiv 2510.17500 v2 pith:SPWGD6LB submitted 2025-10-20 math.NA cs.NA

Paola Goatin , Simone Göttlich , Fabian Ziegler This is my paper · ORCID

classification math.NAcs.NA MSC 35L6565M12

keywords nonlocalconservationlawsheterogeneousmaterialflowconveyorbeltsRoeschemedimensionalsplittingentropysolutionboundedvariationboundaryconditions

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 extends a scalar nonlocal model of conveyor-belt particle transport to a system of N interacting classes in two space dimensions, with walls and obstacles folded into the nonlocal convolution. The central result is that, for bounded-variation initial data and a smooth approximation of the Heaviside activation in the velocity, the Cauchy problem has a unique entropy solution for every finite time horizon, obtained as the limit of approximate solutions built by a Roe scheme with dimensional splitting. The proof supplies explicit L1, L∞, BV and time-compactness bounds, and a Lipschitz continuous dependence on initial data. A sympathetic reader would care because this is the first well-posedness and convergence proof for the multi-class, obstacle-aware version of the model, and the numerical tests reproduce qualitative mixing effects seen in microscopic particle simulations.

What carries the argument

The core machinery is the modified Roe finite-volume scheme with dimensional splitting (Algorithm 3.1), combined with the representation of obstacles as an additional fictitious density class that enters the nonlocal convolution. The dynamic velocity uses a smooth approximation of the Heaviside function with Lipschitz constant L_H; this constant appears in the CFL condition, the L∞ and BV bounds, and the Lipschitz stability estimate, and all estimates degrade as L_H grows. The discrete entropy inequality of Lemma 3.8 upgrades consistency to convergence toward the unique entropy solution.

What would settle it

Simulate a two-class conveyor-belt configuration with initial density locally exceeding r_max and take a sequence of Heaviside approximations with increasing steepness; if the approximate solutions fail to converge, or violate the predicted BV bounds, or the time step must shrink to zero, the central well-posedness result for the smoothed family would be contradicted.

Watch

Extended reading notes

Core claim

The discovery is that the discontinuous Heaviside activation in the dynamic velocity can be replaced by a smooth approximation, and that the resulting regularized system admits a unique entropy weak solution whose discrete approximations are controlled by estimates that are explicit in the smoothing parameter. The convergence argument is carried by a Roe-type scheme with dimensional splitting: the nonlocal velocity is computed once per time step, the flux is split into static and dynamic parts, and the BV and entropy estimates are proved directly on the discrete level. The paper also proves that the solution map is Lipschitz in L1 with respect to initial data, and that the numerical scheme i

Load-bearing premise

The load-bearing assumption is that the Heaviside activation is smoothed into a Lipschitz function; every estimate in the paper scales with the size of its derivative, so if the original discontinuous Heaviside is used, the argument—and the theorem—no longer applies.

Editorial extensions

If this is right

Initial data in (L∞∩BV) produce a unique entropy solution for every T>0, with explicit L1 conservation, L∞ growth, and BV growth bounds.
The Roe scheme with dimensional splitting converges to the entropy solution, giving a practical numerical method for multi-class belt flow with obstacles.
The solution depends Lipschitz continuously on the initial data, so small measurement errors in the initial density lead to controlled errors at later times.
The model, when smoothed, reproduces class-dependent mixing and overtaking effects that match microscopic simulations, suggesting it captures the essential physics of size-heterogeneous cargo.

Reading between the lines

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

Because the well-posedness proof requires a smooth Heaviside approximation and the maximal-density constraint is lost, the theorem does not cover the original discontinuous model; a limit passage as L_H→∞ would be needed to close that gap.
The explicit dependence of the CFL and bounds on L_H suggests an adaptive time-stepping strategy: as the smoothing sharpens, the step size must shrink, which could make the scheme expensive for near-discontinuous activations.
The obstacle-as-fictitious-density construction is a reusable device: any region that should be impermeable can be encoded as a high-density class, potentially extending to multi-domain or moving-obstacle settings.
The same Roe-based framework might be transferred to other nonlocal multi-class models in two dimensions, such as pedestrian or traffic flows, where the flux discontinuity arises from a threshold activation.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The real contribution here is the extension of the scalar Roe-scheme analysis from [14] to an N-class system in 2D, with obstacles handled through an augmented density. The coupled BV/L∞ estimates and the technical Appendix A are the substance, and they look largely correct. The Roe scheme is genuinely less diffusive than the Lax-Friedrichs approach in [1,9], so this has practical value for conveyor belt modeling. I also give credit for being explicit that the smoothing of H is needed and that the maximal density constraint is lost.

The main soft spot is the one the stress-test flagged: Theorem 2.2 is proved under Assumption (H), where the discontinuous Heaviside of the original model (1.2) is replaced by a smooth approximation. Every constant—CFL, L∞ growth, BV bounds—blows up as L_H → ∞, and there is no limit argument back to the original problem. The abstract's claim about treating "the discontinuity in the flux function" overstates what is actually proved. That is a scope gap, not an internal inconsistency, but it is load-bearing for the paper's central claim.

The second issue is that Lemma 3.8, the discrete entropy inequality that underpins the convergence proof, is stated with "the proof is omitted." For a convergence paper this is the largest hole. It may well be analogous to [1,2], but a referee needs to see the argument, especially because the Roe flux with dimensional splitting and the boundary terms are exactly where the analogy can break.

Third, Proposition 3.9 (Lipschitz dependence and uniqueness) leans on Theorem 2.4 of [9], a self-authored paper listed as "to appear." That is not independently checkable. The proof here is only a verification of assumptions, so if [9] changes, this paper changes. Also, the constant K^c(t) is left in the abstract form e^{∫K^c(s)ds} without explicit dependence on the norms, which weakens the claimed quantitative estimate.

Minor: numerical tests are qualitative, no code or data are shipped, and the regularization in the tests is a fixed arctan rather than a sequence converging to H. That limits what the numerics say about the original model.

The paper is worth refereeing. The estimates in Appendix A are elaborate and likely correct, and the system extension is a natural next step. But before acceptance I would want Lemma 3.8 proved or precisely referenced, the scope of the smoothing stated honestly in the abstract and conclusion, and the dependence on [9] resolved. I would not cite it in my own work until those pieces are in place, but I would bring it to a reading group: it is a good example of how far the standard compactness framework pushes, and where the limits are.

Referee Report

4 major / 5 minor

Summary. The paper studies a two-dimensional system of nonlocal conservation laws modeling heterogeneous material flow on conveyor belts, with N particle classes, obstacles encoded via an augmented density, and a dynamic velocity activated by a Heaviside function H. The main result, Theorem 2.2, asserts existence, uniqueness, and BV/L1 estimates for entropy weak solutions obtained as limits of approximate solutions generated by a Roe finite-volume scheme with dimensional splitting (Algorithm 3.1). The proofs rely on positivity, L1, L∞, BV-in-space-time, and discrete entropy estimates, plus a Lipschitz continuous dependence result imported from a companion paper [9]. The numerical section compares two configurations (small/large particle ordering) with microscopic simulations. The central claim is proven only under Assumption (H), which replaces the discontinuous Heaviside by a smooth approximation with bounded derivative L_H, and no limit is established as the smoothing vanishes; the paper explicitly acknowledges that the maximal density constraint is lost.

Significance. If the results hold as stated, the paper is a useful extension of the scalar nonlocal model in [14] to a system of multiple particle classes with obstacles, giving explicit a priori bounds and a finite-volume convergence framework. The technical estimates in Appendix A are elaborate and appear carefully derived. The numerical experiments provide a qualitative validation of the model's behavior. However, the scope of the mathematical theorem is narrower than the abstract and conclusion suggest: well-posedness is proved for a family of smoothed Heaviside models, not for the original discontinuous model (1.1)-(1.2). Two load-bearing ingredients are also deferred: the discrete entropy inequality (Lemma 3.8) is stated without proof, and uniqueness/stability (Proposition 3.9) is imported from an unpublished 'to appear' source. These gaps make the central claim conditional rather than fully established in the submitted manuscript.

major comments (4)

[§2, Assumption (H); Theorem 2.2; §1, last paragraph] The original model (1.1)-(1.2) uses the discontinuous Heaviside H, but Assumption (H) replaces it by a smooth approximation with derivative bounded by L_H. Theorem 2.2 is therefore proved only for the smoothed model. Every quantitative estimate depends on L_H: the CFL condition (3.13), the L∞ growth constant C_c^∞ in (3.12), the BV constants K_1^c, K_2^c in (3.16)-(3.17), and the Lipschitz constant in Proposition 3.9. As the smoothing approaches the true Heaviside, L_H → ∞, so Δt → 0 and the a priori bounds blow up; no compactness argument is given to pass to a solution of the discontinuous model. Section 1 explicitly states that the smoothing is needed for stability and that the maximal-density constraint is lost. Thus the central claim does not cover the model as originally formulated. This is a scope gap rather than an internal inconsistency, but it is load-bearing for the abstract's
[§3.5, Lemma 3.8] The discrete entropy inequality is a load-bearing component: it is used to pass to the limit and identify the limit as an entropy solution. However, the proof is omitted with the comment that it is 'entirely analogous' to [2, Proposition 2.8] and [1, Lemma 2.8]. The setting here is not identical: the scheme uses dimensional splitting, a system of N equations, and a nonlocal flux with an additional convolution through H. A rigorous derivation, or at least a precise statement that the cited result applies verbatim to this setting (including the treatment of the extra sgn terms involving v_stat and J), should be provided. Without it, Theorem 2.2's existence claim is incomplete.
[§3.6, Proposition 3.9] Uniqueness in Theorem 2.2 rests entirely on Proposition 3.9, whose proof invokes 'Theorem 2.4 of [9]', a self-authored paper listed as 'to appear' and hence not independently verifiable by the reader. The manuscript does not state the assumptions (ν) and (J) of that theorem nor reproduce its statement, so the reader cannot check that the present setting satisfies them. Since uniqueness is part of the central well-posedness claim, this is a substantive gap. The authors should either include a self-contained proof of the Lipschitz dependence (or a full statement of the external result with all hypotheses) or make the dependence on the unpublished work explicit in the abstract/introduction.
[§4.2, equation (4.1)] The numerical tests use a fixed arctan approximation of the Heaviside with parameter 50. This is a single regularized model, not a sequence of approximations converging to the discontinuous H. The numerical section therefore cannot serve as evidence for the behavior of the original model, nor does it test the range of validity of the estimates as L_H grows. If the scope of the paper is the regularized model, this is acceptable, but it should be stated clearly; if the intent is to approximate the original model, convergence in the regularization parameter should be examined numerically.

minor comments (5)

[§3.3, Lemma 3.4] The statement says 'for all c∈{0,...,N}' but the proof fixes c∈{1,...,N}; the index set should be {1,...,N}.
[§3.4, Proposition 3.5] The proof defines K_1^c and K_2^c in (3.31)-(3.32) and (3.33)-(3.34), while the theorem statement refers to (3.16)-(3.17). The constants are consistent only after the inequalities 'K_1^c, K_3^c ≤ K_1^c; K_2^c, K_4^c ≤ K_2^c' are used, but the notation is confusing. Please align the definitions or explain the relationship.
[§3.4, equation (3.28)] In the last inequality of (3.28), the term |ρ^c_{i,j}+ρ^c_{i-1,j}| appears; based on the preceding lines it should be |ρ^c_{i,j}-ρ^c_{i-1,j}|. Please correct the typo.
[§4, Figures 2 and 3] Figure 2 shows time values t=1,2,3 while Figure 3 shows t=100,200,300. The reason for this difference is not explained; if the two cases use different time scales or parameter regimes, this should be stated in the text.
[Appendix A, (A.7)-(A.11)] The notation 'k' is used in the convolution sums (e.g., ∂_1η_c(x_{i+1/2}-k, y_j-ℓ)) where the summation index h was introduced; this is a minor typo that should be fixed.

Circularity Check

1 steps flagged · score 4.0 of 10

Uniqueness/stability is delegated to an unverified same-author 'to appear' theorem; existence/convergence is otherwise derived in-paper.

uniqueness imported from authors [Section 3.6, Proposition 3.9 (proof), used in Theorem 2.2 for uniqueness]
"For this proof we rely on a result in [9], where the stability with respect to the initial data was shown under less restrictive assumptions. To show the Lipschitz estimate with respect to the initial data it is only necessary to verify the assumptions (ν) and (J) of [9, Theorem 2.4]."

The uniqueness part of Theorem 2.2 is explicitly delegated to Proposition 3.9, whose proof only verifies hypotheses of [9, Theorem 2.4] — a 'to appear' paper with an overlapping author (Goatin). No derivation of the Lipschitz estimate is given here; the estimate K^c(t) is quoted from [9]. The logical chain 'Theorem 2.2 uniqueness ⇐ Prop. 3.9 ⇐ [9, Thm 2.4]' thus terminates in an unverified same-author citation rather than in an argument contained in, or independently checkable from, this paper. This is load-bearing self-citation for a central component of the advertised well-posedness theorem, although the existence/convergence part is derived in Sections 3.1–3.5.

full rationale

Most of the derivation is self-contained: positivity, L^1, L^∞, BV-in-space/time estimates are proven directly from the Roe scheme (Lemmas 3.2–3.4, Proposition 3.5, Corollary 3.7), and the constants depend on the data and on L_H. The discrete entropy inequality is quoted from published [2]/[1] as a standard monotone-scheme argument, not from a fitted quantity, so it does not create circularity by itself. The numerical tests compare to microscopic simulations rather than fitting the macroscopic model to its own output, so no 'fitted input called prediction' pattern is present. The one load-bearing circularity-adjacent issue is Proposition 3.9: uniqueness/stability in Theorem 2.2 is not proved in this paper but imported from [9, Theorem 2.4], a same-author paper 'to appear' that the reader cannot independently check. That is self-citation used as the terminal justification of a central claim. Separately, the paper explicitly acknowledges that the original discontinuous Heaviside is replaced by a smooth approximation and that the constants and CFL condition (3.13) blow up with L_H = ||H'||_{L∞}; no limit as the regularization vanishes is established. This is an honest scope gap between the original model (1.2) and Theorem 2.2, not a definitional circularity. Overall, because the existence and approximation result is independently derived while uniqueness is delegated to an unreviewable same-author theorem, the circularity score is 4.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The proof is a standard compactness argument, so no parameters are fitted to the conclusion; all bounds are explicit in terms of data norms. The main external input is the self-authored 'to appear' theorem [9] for the stability part. The smoothing of the Heaviside is a central modeling assumption that prevents the result from covering the original discontinuous model.

assumptions (4)

domain assumption v_stat ∈ C^2(R^2;R^2), mollifiers η_c, η̃ ∈ (C^3 ∩ W^{3,∞})(R^2;R_+), H smooth approximation of Heaviside with bounded derivative L_H.
Assumptions (v), (η), (H), Section 2. These regularity assumptions make the velocity field C^2 and enable the BV estimates; the boundedness of L_H is required in the CFL condition (3.13) and all bounds.
domain assumption Inward flow condition (v_stat + v_dyn_c)·n ≤ 0 on ∂Ω, ensured by choosing artificial densities R_{N+l} large enough.
Assumptions (Ω.1)-(Ω.2) and construction (1.3). The existence of sufficiently large R_{N+l} is assumed to make the boundary impenetrable; no quantitative bound is given.
ad hoc to paper [9, Theorem 2.4] — well-posedness and Lipschitz dependence for a multi-population nonlocal pedestrian model.
Prop 3.9 (and hence uniqueness in Theorem 2.2) is obtained by verifying assumptions of this theorem from a self-authored paper marked 'to appear' in the references; it is not independently accessible.
domain assumption The discrete convolution quadrature (3.7) is a consistent approximation of the continuous convolution; quadrature error is not accounted for in the convergence proof.
Algorithm 3.1 and equation (3.7). The scheme uses a Riemann-sum quadrature for ∂_i η_c * r. Lemma A.1 bounds only differences of these quadrature sums; convergence to the continuous convolution as Δx,Δy→0 is not explicitly proven.

invented entities (1)

Artificial obstacle densities R_{N+l} χ_{Ω_c^l} in the augmented density r_Ω
purpose: Enforce the tangential/inward boundary condition by placing large artificial densities outside the physical domain Ω, so that objects do not cross walls.
Equation (1.3). This is a computational bookkeeping device from refs [4,8,9], not a measurable physical quantity; it has no falsifiable handle outside the model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A nonlocal model for heterogeneous material flow on conveyor belts." pith.science (2026). https://pith.science/paper/SPWGD6LB

@misc{pith2026251017500,
  author       = {Pith},
  title        = {Pith review of: A nonlocal model for heterogeneous material flow on conveyor belts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SPWGD6LB}},
  note         = {Machine review of arXiv:2510.17500}
}

read the original abstract

In this paper, a finite volume approximation scheme is used to solve a nonlocal macroscopic material flow model in two space dimensions, accounting for the presence of boundaries in the nonlocal terms. Based on a previous result for the scalar case, we extend the setting to a system of heterogeneous material on bounded domains. We prove the convergence of the approximate solutions constructed using the Roe scheme with dimensiona splitting, where the major challenge lies in the treatment of the discontinuity occurring in the flux function. Numerical tests show a good agreement with microscopic simulations.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Monotone-based Numerical Schemes for Two-Dimensional Systems of Nonlocal Conservation Laws
math.NA 2026-01 conditional novelty 6.0 of 10

Monotone finite-volume schemes with approximated nonlocal terms converge to the unique weak entropy solution of 2D nonlocal conservation-law systems, at worst-case rate O(√Δt).

Reference graph

Works this paper leans on

15 extracted references · cited by 1 Pith paper

[14]

Rossi, J

E. Rossi, J. Weißen, P. Goatin, and S. G¨ ottlich. Well-posedness of a non-local model for material flow on conveyor belts. ESAIM Math. Model. Numer. Anal., 54(2):679–704, 2020

2020
[9]

Goatin and E

P. Goatin and E. Rossi. Well-posedness of nonlocal macroscopic models of multi-population pedestrian flows for domain shape optimization. J. Hyperbolic Differ. Equ., to appear
[1]

Aggarwal, R

A. Aggarwal, R. M. Colombo, and P. Goatin. Nonlocal systems of conservation laws in several space dimensions. SIAM J. Numer. Anal., 53(2):963–983, 2015

2015
[2]

Amorim, R

P. Amorim, R. M. Colombo, and A. Teixeira. On the numerical integration of scalar nonlocal conservation laws. ESAIM Math. Model. Numer. Anal., 49(1):19–37, 2015

2015
[3]

Blandin and P

S. Blandin and P. Goatin. Well-posedness of a conservation law with non-local flux arising in traffic flow modeling. Numer. Math., 132(2):217–241, 2016

2016
[4]

B¨ urger, P

R. B¨ urger, P. Goatin, D. Inzunza, and L. M. Villada. A non-local pedestrian flow model accounting for anisotropic interactions and domain boundaries. Math. Biosci. Eng., 17(5):5883–5906, 2020

2020
[5]

R. M. Colombo, M. Garavello, and M. L´ ecureux-Mercier. A class of nonlocal models for pedestrian traffic. Math. Models Methods Appl. Sci., 22(4):1150023, 34, 2012. 28

2012
[6]

Crandall and A

M. Crandall and A. Majda. The method of fractional steps for conservation laws. Numer. Math., 34(3):285–314, 1980

1980

Show all 15 references

[7]

M. G. Crandall and A. Majda. Monotone difference approximations for scalar conservation laws. Math. Comp., 34(149):1–21, 1980

1980
[8]

Goatin, D

P. Goatin, D. Inzunza, and L. M. Villada. Nonlocal macroscopic models of multi-population pedestrian flows for walking facilities optimization. Appl. Math. Model., 141:Paper No. 115927, 14, 2025

2025
[10]

G¨ ottlich, S

S. G¨ ottlich, S. Hoher, P. Schindler, V. Schleper, and A. Verl. Modeling, simulation and validation of material flow on conveyor belts. Appl. Math. Model., 38(13):3295–3313, 2014

2014
[11]

Keimer and L

A. Keimer and L. Pflug. Existence, uniqueness and regularity results on nonlocal balance laws. J. Differential Equations, 263(7):4023–4069, 2017

2017
[12]

Keimer, L

A. Keimer, L. Pflug, and M. Spinola. Existence, uniqueness and regularity of multi-dimensional nonlocal balance laws with damping. J. Math. Anal. Appl., 466(1):18–55, 2018

2018
[13]

S. N. Kruˇ zhkov. First order quasilinear equations with several independent variables. Mat. Sb. (N.S.), 81 (123):228–255, 1970

1970
[15]

Van Lint, S

J. Van Lint, S. P. Hoogendoorn, and M. Schreuder. Fastlane: New multiclass first-order traffic flow model. Transportation Research Record, 2088(1):177–187, 2008. 29

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [14]

Rossi, J

E. Rossi, J. Weißen, P. Goatin, and S. G¨ ottlich. Well-posedness of a non-local model for material flow on conveyor belts. ESAIM Math. Model. Numer. Anal., 54(2):679–704, 2020

2020

[2] [9]

Goatin and E

P. Goatin and E. Rossi. Well-posedness of nonlocal macroscopic models of multi-population pedestrian flows for domain shape optimization. J. Hyperbolic Differ. Equ., to appear

[3] [1]

Aggarwal, R

A. Aggarwal, R. M. Colombo, and P. Goatin. Nonlocal systems of conservation laws in several space dimensions. SIAM J. Numer. Anal., 53(2):963–983, 2015

2015

[4] [2]

Amorim, R

P. Amorim, R. M. Colombo, and A. Teixeira. On the numerical integration of scalar nonlocal conservation laws. ESAIM Math. Model. Numer. Anal., 49(1):19–37, 2015

2015

[5] [3]

Blandin and P

S. Blandin and P. Goatin. Well-posedness of a conservation law with non-local flux arising in traffic flow modeling. Numer. Math., 132(2):217–241, 2016

2016

[6] [4]

B¨ urger, P

R. B¨ urger, P. Goatin, D. Inzunza, and L. M. Villada. A non-local pedestrian flow model accounting for anisotropic interactions and domain boundaries. Math. Biosci. Eng., 17(5):5883–5906, 2020

2020

[7] [5]

R. M. Colombo, M. Garavello, and M. L´ ecureux-Mercier. A class of nonlocal models for pedestrian traffic. Math. Models Methods Appl. Sci., 22(4):1150023, 34, 2012. 28

2012

[8] [6]

Crandall and A

M. Crandall and A. Majda. The method of fractional steps for conservation laws. Numer. Math., 34(3):285–314, 1980

1980