{"id":"74df3fad-30db-4efc-9614-74b9947a065b","arxiv_id":"2508.16333","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A state-dependent sweeping process models elastoplastic spring lattices with softening, hardening and perfect plasticity, yielding non-unique solutions and shear-band localization.","lead":"This paper rewires the mechanics of lattices of elasto-plastic springs, including softening springs, as a state-dependent sweeping process, a moving-constraint evolution equation. The framework reproduces known non-uniqueness in a two-spring toy model and produces asymmetric shear bands in larger lattice simulations.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No local existence theorem for softening leaves the Section 6 shear-band simulations without a guarantee that a solution of (58)–(59) exists; the failure of condition (93) is acknowledged in Remark 4.5 and deferred in Section 7.","rationale":"The reader's weakest assumption is exactly where I would also locate the load-bearing issue. Theorem 4.1 is a well-documented equivalence and its proof is the strongest part of the paper; the single-spring analysis and the reduced-dimension formulation are also convincing. The numerical examples are reproducible in principle (code is supplied) and the shear-band phenomenology matches the engineering literature, which gives independent support. However, the softening regime is the paper's claimed novelty, and for that regime there is no existence theorem. Remark 4.5 concedes that the only standard sufficient condition fails, Remark 4.6 concedes that Algorithm 1 can diverge when only unstable fixed points exist, and Section 7 explicitly lists a local existence theorem as future work. Without such a theorem, the fixed points found by iteration are not known to be approximations of solutions of (58)–(59). This is not a mathematical inconsistency in the paper, but it is a genuine gap in the support for the headline claim that the sweeping process 'solves the evolution'. The secondary issue—Proposition 5.1's proof is only sketched—reinforces the need for verification, but I would not make it the primary attack because the explicit branches and the comparison with [CB04] make the non-uniqueness claim plausible, and the omitted algebra can be checked independently. A symbolic/global fixed-point check on the two-spring example would settle the existence question in that concrete case; if it succeeds, the remaining gap is confined to larger lattices, where a local existence theorem is still needed. The appropriate verdict remains CONDITIONAL: the mathematical skeleton is strong, but the softening existence statement must be completed before the numerical results can be accepted as a solution theory.","tokens_in":34305,"tokens_out":5927,"duration_ms":71271,"concrete_test":"Choose the two-spring softening system of Section 5 with the parameters of Fig. 8(c), and at each time step before complete failure solve the fixed-point equation (91) exactly, not by iteration: represent C(z,t) by its facet inequalities (103)–(104), and use branch-and-bound or a linear-complementarity solver to list all fixed points. Then repeat with the time step halved. If at any step the fixed-point set is empty, or if the set changes discontinuously with Δt, then the existence-gap concern lands and the numerical branches in Fig. 8 are not guaranteed to correspond to solutions of (58)–(59). If nonempty and consistent, repeat the exact fixed-point solve on a small rectangular lattice near the reported symmetry-breaking time in Section 6.1.1; an empty set there would invalidate the shear-band claim, while a nonempty set would support it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing gap is existence of solutions to the state-dependent sweeping process (58)–(59) in the softening regime used for the paper's main numerical claim. Theorem 4.1 proves equivalence between (LSM1)–(LSM6) and (58)–(59) only for a given solution of the latter; it does not supply existence. The standard existence theorem (Theorem 1.2) requires the Lipschitz condition (93), and Remark 4.5 explicitly notes that (93) fails when β_i ≤ 1, i.e. exactly for perfectly plastic and softening springs. Algorithm 1/2 therefore advances by searching for fixed points of (91), and Remark 4.6 states that the algorithm diverges when no fixed point or only unstable fixed points exist. Section 7 defers a local existence theorem to future work. Consequently the Section 6 shear-band plots and the bifurcation inventory in Fig. 8 are not yet known to be trajectories of (58)–(59); they are fixed points of an iterated projection found numerically. This is an explicitly acknowledged limitation rather than a contradiction, but it is the central support for the softening part of the advertised claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a state-dependent sweeping process formulation for quasi-static elasto-plastic lattice spring models whose springs may harden, soften, or be perfectly plastic. After analyzing a single spring (§3), it formulates the full lattice equations (LSM1)–(LSM6) and proves in Theorem 4.1 an equivalence between these equations and the state-dependent sweeping process (58)–(59), including reconstruction formulas (60)–(65). It then introduces implicit catch-up algorithms (§4.3–4.4), analyzes a two-spring toy model (§5) where non-uniqueness is asserted in Proposition 5.1, and presents numerical simulations of rectangular and triangular lattices (§6) showing shear-band localization in softening springs. The author is explicit that for softening and perfect plasticity the standard existence theorem (Theorem 1.2) does not apply because condition (93) fails, and that a local existence theorem is future work (§7).","tokens_in":34664,"tokens_out":5786,"duration_ms":66257,"significance":"If the central equivalence and the existence gap were closed, the paper would provide a unified finite-dimensional framework for rate-independent plasticity with softening, connecting the mechanical problem to a state-dependent sweeping process and enabling a rigorous study of non-uniqueness and shear-band bifurcations. The two-sided equivalence proof of Theorem 4.1 is detailed and self-contained, the toy-model analysis addresses a genuine non-uniqueness phenomenon, and the numerical experiments are accompanied by available code and videos. The derivation does not appear to presuppose its conclusions; the main weakness is not circularity but a missing existence theorem in the softening regime. As it stands, the advertised claim that the sweeping process 'solves' the evolution with softening outruns the proved results.","major_comments":[{"comment":"The paper's central claim is that the state-dependent sweeping process solves the evolution for arbitrary mixtures including softening, but no existence theorem is proved in this regime. Theorem 1.2 requires the Lipschitz condition (93), and Remark 4.5 explicitly notes that (93) cannot hold when β_i ≤ 1 (softening/perfectly plastic springs). Section 7 defers a local existence result to future work. Since Theorem 4.1 only establishes equivalence for a given solution of (58)–(59), and Algorithm 1/2 is defined by fixed points of (91)/(99), the Section 6 simulations are currently not known to be solutions of the time-continuous problem. Please either prove existence of a fixed point for each time step under a suitable non-degeneracy assumption, or explicitly restrict the claims to the discrete fixed-point problem and state the absence of existence as a limitation in the abstract and conclusi","section":"§4.2, Remark 4.5, §7"},{"comment":"The analytical non-uniqueness result is only sketched. The proof says 'To save space we will omit complete derivations' and reports that equations were solved manually. Formulas (108)–(109) and the case analysis are central to the claim of coexisting solutions and to the bifurcation picture in Figure 8. Please provide a complete derivation (or put it in a supplementary file) with the inequality checks (110) and all case distinctions, rather than a sketch.","section":"§5.3, Proposition 5.1"},{"comment":"The numerical experiments do not report the time-step size, the tolerance used in Algorithm 2, the maximum number of iterations, or the initial-guess policy; Remark 4.6 states that iteration may diverge when there are no stable fixed points. For the shear-band runs in Figs. 10–16, please report for every run that the convergence criterion was met at each time step, and provide data underlying the iteration-count plots. This is necessary to distinguish genuine solutions from algorithm failure and to support the claim that modifying the initial guess yields a second solution (Fig. 3 vs Fig. 10d).","section":"§6, Algorithm 2"},{"comment":"In the sweep-to-LSM direction the proof says 'we do not consider the state of complete failure' and uses implication (41). The theorem statement, however, does not state this exclusion. If the equivalence is intended to cover complete-failure states, the proof is incomplete; if not, the statement and the abstract's 'arbitrary placement' should be qualified. This is not merely cosmetic, since complete failure is reached in the softening simulations (Remark 3.1).","section":"§4.2, Theorem 4.1 proof"}],"minor_comments":[{"comment":"Several typos and language issues: 'Given finctions' should be 'Given functions'; 'not not alter' in §4; 'modyfing' and 'incuded' in §6. A careful proofreading pass is needed.","section":"§1, §4.1.2, §6"},{"comment":"The notation for the moving set is very dense. A short table collecting the symbols in (55), (60), (70), (72), (73)–(74) would improve readability.","section":"§4.2, equations (67)–(74)"},{"comment":"The captions do not define the coordinate axes or the meaning of the wedges and red/black points. Please add explicit axis labels and a legend.","section":"§5, Figures 8–9"},{"comment":"The numerical setup is incomplete: no time-step size, tolerance, or iteration limits are given. These should be stated for reproducibility, especially because Algorithm 2's convergence is not guaranteed in the softening case.","section":"§6"},{"comment":"The paper relies heavily on [Gud+23] for definitions and algebraic properties. Please state explicitly in Section 4 which results are taken from that reference and which are new, so that the equivalence theorem can be checked without reading the full prior paper.","section":"§4.1"}],"recommendation":"major_revision","confidential_remarks":"The author is transparent about the missing existence theorem and explicitly lists it as future work in §7, so this is not a hidden flaw. The problem is that the paper's framing ('solves evolution ... with softening') exceeds what is proved. I would not recommend reject, because the manuscript could be made acceptable either by adding a local existence/fixed-point result under non-degeneracy or by carefully narrowing the claims and emphasizing the conditional nature of the numerical results. The equivalence theorem is a solid contribution, and the numerical shear-band experiments are valuable even if presently only discrete fixed-point computations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: the paper is worth a referee's time, but the referee should push hard on the existence gap. What's new: Theorem 4.1 gives a unified state-dependent sweeping process for lattice spring models with softening, hardening, and perfect plasticity in arbitrary mixture. The proof is detailed and, as far as I can tell, correct. For the first time we get a finite-dimensional formulation of rate-independent plasticity with softening in a mechanical model, and the numerical section shows shear band localization and non-uniqueness, which matches engineering intuition. The author supplies code and videos, which is good practice.\n\nSoft spots, in proportion: The big one is that the softening regime has no existence theorem. The standard conditions for the state-dependent sweeping process fail when beta_i <=1 (Remark 4.5), and Theorem 1.2 only covers hardening. The author says a local existence result is future work (Section 7). That means the Section 6 shear bands are not yet known to be trajectories of (58)-(59); they are fixed points of an iterated projection found numerically. That's not a fatal flaw—the author is transparent about it—but it is the load-bearing support for the paper's advertised claim, and it needs an answer before the result can be called a solution theory.\n\nProposition 5.1's proof is sketched ('idea of the proof'), with algebra omitted. In a paper whose main analytical novelty is non-uniqueness, that's a smaller but real gap. The numerical section also lacks sensitivity analysis: no parameter sweeps, no mesh convergence study, just one example per parameter set. That's fine for a demonstration, but it limits how much can be concluded about the 'emergence' of shear bands.\n\nThe citation pattern is honest: the non-uniqueness in two springs was already in Chen & Baker 2004 and the author says so. The reliance on the author's earlier lattice construction [Gud+23] is appropriate, not a red flag.\n\nWho gets value: specialists in sweeping processes, plasticity, and lattice models. A serious referee can handle this. I would send it to review, with instructions that the existence gap be addressed either by a theorem or by a sharp restriction of the claims.","headline":"A genuinely useful sweeping-process formulation for lattice elastoplasticity with softening, but the paper's main numerical claim rests on an unproven existence step that the author explicitly defers.","tokens_in":35060,"tokens_out":1955,"would_cite":true,"duration_ms":21496,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74C05","74N30","47J20","47J26"],"pacs":[],"model":"deepseek-v4-flash","headline":"The full Lattice Spring Model with softening, hardening, and perfectly plastic springs is equivalent to a state-dependent sweeping process.","keywords":["state-dependent sweeping process","elastoplasticity with softening","non-uniqueness of solutions","shear band","strain localization","lattice spring model","implicit catch-up algorithm","plasticity bifurcation"],"falsifier":"Take the two-spring softening system of Section 5, set both springs to softening parameters with no complete failure, and choose the initial condition on the distributed saddle solution satisfying (106)-(107). Run the implicit catch-up iterations with a small time step; if for any step before failure the projection map has no fixed point or the iteration diverges or cycles, then the claimed sweeping-process formulation does not produce the evolution asserted for softening.","tokens_in":34210,"feed_emoji":"🧱","tokens_out":7400,"duration_ms":78171,"temperature":0.7,"pith_summary":"This paper tries to show that the quasi-static evolution of a lattice of elasto-plastic springs—including the mathematically troublesome softening case—can be written exactly as a state-dependent sweeping process, a differential inclusion with a moving convex set. If the equivalence is right, the same numerical and analytical machinery applies to hardening, perfect plasticity, and softening, with the difference between them reduced to a parameter. The paper demonstrates the payoff by analytically exhibiting coexisting solutions in a two-spring softening system and by simulating shear-band formation with strain localization in regular lattices. This matters because softening plasticity is normally ill-posed: stress-strain relations lose monotonicity, models become mesh-dependent, and solutions can branch. A single finite-dimensional formulation that still produces well-defined evolutions up to spring failure would give a rigorous route into ductile-fracture initiation.","feed_headline":"One sweeping process covers softening, hardening, perfect plasticity","feed_subtitle":"The formulation turns ill-posed softening plasticity into a tractable inclusion, and shear bands emerge in simulations.","key_machinery":"The load-bearing object is the state-dependent moving set C(y, q): for each spring, a convex polytope of admissible stresses whose upper and lower thresholds shift with the damage variable q and with the applied displacement load. The sweeping-process inclusion dot-y in -N_C(y,q)(y), together with the linear compatibility and equilibrium constraints, encodes the whole lattice. The numerical scheme is the implicit catch-up map, whose fixed points at each time step are the discrete solutions; stability of these fixed points under iteration is what the paper uses to explain bifurcation and branching in softening.","core_discovery":"At the center is a finite-dimensional equivalence (Theorem 4.1): the full displacement-stress-damage system (LSM1)-(LSM6) for a lattice built of springs of any mixture—hardening, perfectly plastic, softening—holds if and only if the pair (y, q) of reduced stress and damage variables solves the state-dependent sweeping process (58)-(59), with the moving set C(y, q) the polytope of admissible stresses. The equivalence is two-sided, which is stronger than earlier sweeping-process formulations for perfectly plastic lattices. For the toy lattice of two springs in series, the paper solves the sweeping process explicitly and finds coexisting solutions when both springs soften: a distributed solutio","pith_inferences":["If the promised local existence theorem for softening is established, the discrete model could serve as a regularized testbed in which shear-band width is controlled by lattice spacing, allowing mesh-dependence questions to be studied by refinement.","The observed relation between fixed-point stability and energy minima suggests a conjecture the paper does not prove: each iteration of the catch-up map decreases an underlying energy toward a local minimum; a cycling example would refute it, and a proof would unify variational and sweeping-process approaches.","The degeneracy in the plasticity-modulus formula means many parameter pairs produce the same observable modulus, so calibration from macroscopic stress-strain curves alone cannot identify them; extra physical or atomistic information would be needed to predict damage evolution and failure location.","A measure-valued state-dependent sweeping process for continuous media, which the paper sketches as conceivable, would complement the existing continuum variational theory for softening and could yield two-sided error estimates and optimal-control formulations for ductile-fracture initiation."],"forward_implications":["Changes in the plastic regime become parameter changes: beta crossing 1 turns a hardening spring into a softening one, so one solver and one theory cover all three cases.","Softening lattices admit multiple coexisting evolutions; the catch-up iteration's fixed points classify them as stable or saddle, giving a route to predict which branch a load path selects.","Shear bands with strain localization appear as emergent, symmetry-breaking events in symmetric lattices, and their orientation is controlled by lattice geometry and defects, so the model can reproduce ductile-fracture precursors.","Evolution can be computed reliably only up to the first spring's complete failure; past that point the sweeping process has no admissible continuation, matching the idea of material separation.","The two-spring toy model provides a minimal analytically solvable case where the nonsmooth bifurcation underlying softening can be studied rigorously."],"supporting_citations":[{"why":"Supplies the state-dependent sweeping process framework and the implicit catch-up algorithm whose fixed points define the numerical scheme.","marker":"[KM98]"},{"why":"States the existence theorem for state-dependent sweeping processes whose Lipschitz condition fails for softening springs, marking the gap the paper addresses numerically.","marker":"[KM00]"},{"why":"Earlier sweeping-process formulation for elastic-perfectly plastic lattice spring models that Theorem 4.1 extends to softening and hardening.","marker":"[Gud+23]"},{"why":"Energy-profile and bifurcation analysis of the same two-spring softening model, used to confirm the non-uniqueness result of Proposition 5.1.","marker":"[CB04]"},{"why":"Mathematical theory of hardening plasticity used as the well-posedness baseline and for elastic-unloading terminology.","marker":"[HR12]"},{"why":"Discusses uniqueness and non-uniqueness in state-dependent sweeping processes, framing why coupled softening elements lose uniqueness.","marker":"[BKS04]"}],"fun_headline_variants":["Sweeping process solution reveals shear bands in spring lattices","State-dependent sweeping process covers softening and hardening springs","Non-unique softening solutions and shear bands from one process","Softening plasticity: a unified sweeping process for all springs"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The implied existence of a well-defined solution for softening lattices: at every time step the implicit catch-up map must have a fixed point that the iteration can actually reach, even though the standard existence condition behind the cited theorem fails exactly for softening springs.","fun_headline_variants_meta":{"raw":{"variants":["Sweeping process solution reveals shear bands in spring lattices","State-dependent sweeping process covers softening and hardening springs","Non-unique softening solutions and shear bands from one process","Softening plasticity: a unified sweeping process for all springs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000272,"raw_usage":{"total_tokens":1437,"prompt_tokens":679,"completion_tokens":758,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":423,"completion_tokens_details":{"reasoning_tokens":701}},"tokens_in":423,"tokens_out":758,"duration_ms":9004,"temperature":1.0,"reasoning_tokens":701,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T17:23:00.724215+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the two-spring softening system of Section 5, set both springs to softening parameters with no complete failure, and choose the initial condition on the distributed saddle solution satisfying (106)-(107). Run the implicit catch-up iterations with a small time step; if for any step before failure the projection map has no fixed point or the iteration diverges or cycles, then the claimed sweeping-process formulation does not produce the evolution asserted for softening.","supporting_citations":[],"review_version":1}