{"id":"2f7eff7e-2532-4364-a2a8-d9d190396f6e","arxiv_id":"1909.00588","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For the spectral fractional Laplacian on a bounded domain, obstacle solutions satisfy the Lewy-Stampacchia inequality f ≤ Au ≤ max{f, Aψ}, which yields well-posedness of a fractional unidirectional diffusion equation.","lead":"The paper proves a two-sided pointwise estimate for obstacle problems driven by the spectral fractional Laplacian in a bounded domain. The estimate is then used to show existence, uniqueness, and stability for a fractional one-way diffusion equation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The time-discrete functional (31) in Lemma 5.1 is not the energy associated with A_σ = (-Δ)^s + 1/τ; as written, its Euler-Lagrange equation is not (30), so the proof of Theorem 1.11 does not follow from Theorem 1.6.","rationale":"The reader's weakest assumption was (12), the regularity of the obstacle's fractional Laplacian. I do not dispute that assumption; it is explicitly imposed and the proof of Theorem 1.6 uses it in a transparent way. My stress-test found no fatal gap in the Lewy-Stampacchia theorem itself. The more load-bearing concern is an internal inconsistency in the application: equation (31) defines a functional that, literally read, is not the energy of A_σ and cannot yield the discrete obstacle problem (30) used to prove Theorem 1.11. This is especially important because the paper's abstract advertises well-posedness of anomalous unidirectional diffusion as a consequence of the Lewy-Stampacchia inequality. The reader had already marked the application half as incomplete because Theorems 1.11-1.13 are sketched or omitted; my concern identifies a specific mathematical mismatch inside the sketch of Lemma 5.1. Since the fix is likely a typographical correction (replace |(-Δ)^s v|^2 by |(-Δ)^{s/2}v|^2), I would keep the verdict conditional rather than reject the paper. I mark disagreement with the reader's identified weakest assumption because the most vulnerable point is elsewhere, while agreeing with the overall conditional verdict.","tokens_in":22788,"tokens_out":32247,"duration_ms":266303,"concrete_test":"Derive the first-order optimality condition for the functional printed in (31) over the closed convex set K_0^k. With the printed quadratic term (1/2)∫|(-Δ)^s v|^2, the variation against admissible w-v contains ∫(-Δ)^s v · (-Δ)^s(w-v), which cannot be rewritten as ⟨(-Δ)^s v, w-v⟩ for arbitrary w ∈ H_0^s and does not yield (30). If instead the intended term is (1/2)∫|(-Δ)^{s/2}v|^2, the variation is exactly ⟨(-Δ)^s v + (1/τ)v, w-v⟩, matching Lemmas 2.1 and 2.5. Checking which expression is used in the subsequent estimates (38)-(46) settles whether Theorem 1.11 has a proof or only a typo.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central Lewy-Stampacchia proof in Theorem 1.6 appears coherent: Lemma 2.2 gives the needed nonlocal sign identity, Lemma 2.3 correctly upgrades K2-elements to X_0^{2s}, and Lemma 2.5 supplies the equivalence with the truncated constraint. The main internal inconsistency I find is in the application half. Lemma 5.1 sets A = A_σ = (-Δ)^s + 1/τ_k and invokes Lemma 2.1/2.5, whose variational functional is J(v) = (1/2)⟨A v, v⟩ - ⟨f, v⟩ = (1/2)‖(-Δ)^{s/2}v‖_{L^2}^2 + (1/(2τ_k))‖v‖_{L^2}^2 - ⟨..., v⟩. But the displayed J_k in (31) uses (1/2)∫_Ω |(-Δ)^s v|^2 dx. For general v ∈ H_0^s, this functional is not even finite, since (-Δ)^s v is in H^{-s}(Ω), not L^2(Ω). If the domain is restricted to X_0^{2s}, the first variation of ‖(-Δ)^s v‖^2 is not ⟨(-Δ)^s v, w-v⟩ but a term involving (-Δ)^{2s}; it does not produce the obstacle equation (30) or the two-sided bound (38). Thus, unless (31) is corrected to |(-Δ)^{s/2}v|^2, the existence proof for the unidirectional diffusion equation is unsupported as written. This is a concrete, fixable defect, not a flaw in Theorem 1.6 itself.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a Lewy-Stampacchia-type inequality for the spectral fractional Laplacian on bounded Lipschitz domains. Under the assumptions f in L^2(Omega) and (-Delta)^s psi being a signed Radon measure whose positive part lies in L^2(Omega), the solution u of the obstacle variational inequality is shown to lie in X_0^{2s}(Omega) and to satisfy f <= Au <= max{f, A psi} a.e. in Omega, where A = (-Delta)^s + lambda. The proof follows Gustafsson's dual formulation, combined with the Caffarelli-Silvestre extension and a nonlocal sign estimate. The paper then applies this inequality to an anomalous unidirectional diffusion equation, claiming uniqueness, stability, existence via implicit Euler time discretization, comparison, and long-time convergence of strong solutions.","tokens_in":23128,"tokens_out":11231,"duration_ms":96094,"significance":"If the main theorem is correct, it constitutes a substantial extension of Lewy-Stampacchia estimates to the spectral fractional Laplacian, including L^2 regularity of (-Delta)^s u, and the application to strong solutions of unidirectional fractional diffusion is new. The proof of Theorem 1.6 appears internally sound: the key lemmas (2.1, 2.2, 2.3, and 2.5) form a coherent chain, the nonlocal sign estimate of Lemma 2.2 is correctly used, and the right-hand side max{f, A psi} is fixed by the obstacle rather than fitted. No free parameters or circular normalizations appear in the derivation. However, the application half contains a concrete error in the time-discrete functional, so the present version does not establish the existence theorem for the anomalous diffusion equation.","major_comments":[{"comment":"The functional J_k displayed in (31) is not the energy associated with the operator A_sigma = (-Delta)^s + 1/tau_k used in the same lemma. For that operator the natural variational functional is (1/2)∫_Ω |(-Delta)^{s/2} v|^2 dx + (1/(2 tau_k))∫_Ω |v|^2 dx - ⟨u_{k-1}/tau_k + f_k, v⟩, whereas (31) contains (1/2)∫_Ω |(-Delta)^s v|^2 dx. The latter is not finite for a general v in H_0^s(Ω), because (-Delta)^s v belongs to H^{-s}(Ω) rather than L^2(Ω), and its first variation is not ⟨A_sigma v, w-v⟩. Consequently the minimizer of (31) does not satisfy the implicit Euler equation (30), and the invocation of Lemmas 2.1 and 2.5 to obtain (34)-(38) is not justified. This is a load-bearing defect for the existence theorem.","section":"Section 5, Eq. (31) and Lemma 5.1"},{"comment":"The claim that there exists a unique u_1 in K_0^1 minimizing J_1 given by (31), and that (34)-(37) follow from Lemmas 2.1 and 2.5, is unsupported as written because J_1 is not the functional considered in those lemmas. The two-sided estimate (38), which is the only point where the Lewy-Stampacchia inequality enters the time-discrete existence proof, is therefore unproved. The defect appears fixable by replacing |(-Delta)^s v|^2 with |(-Delta)^{s/2} v|^2 in (31) and re-checking the subsequent estimates, but as written the proof of Theorem 1.11 is incomplete.","section":"Section 5, proof of Lemma 5.1, Step 1"}],"minor_comments":[{"comment":"If the intended functional is indeed the one associated with A_sigma, the same correction should be propagated consistently through the proof of Lemma 5.1; the notation ‖(-Delta)^s v‖^2 is otherwise ambiguous and suggests the incorrect energy.","section":"Equation (31)"},{"comment":"The phrase 'μ + μζ' appears to be a typo for 'μ + ζ'; please correct it.","section":"Appendix A, Lemma A.1"},{"comment":"The proofs of Theorems 1.12 and 1.13 are omitted on the grounds that they follow from [AK19]; since these results depend on the corrected existence argument, please state explicitly which arguments from [AK19] carry over and what fractional-specific modifications are needed.","section":"Section 1.7"},{"comment":"There are numerous typographical artifacts, e.g., 'unidirectio nal' and 'inequal ity'; a careful proofread is recommended.","section":"Abstract and throughout"}],"recommendation":"major_revision","confidential_remarks":"The core Lewy-Stampacchia proof in Theorem 1.6 appears sound and the paper has merit, but the application section currently contains a clear functional mismatch in Lemma 5.1 that undermines the existence theorem. The defect is local and fixable by replacing the energy term in (31). I recommend requesting a revision, with careful re-verification of the time-discrete argument and the final Minty step, which is only sketched."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main result is a genuine Lewy-Stampacchia inequality for the spectral fractional Laplacian on bounded Lipschitz domains, with pointwise two-sided bounds and X_0^{2s} regularity. The proof via Caffarelli-Silvestre extension and Gustafsson's dual formulation is sound; I checked Lemma 2.2's sign identity and the duality argument in Lemma 2.3, and both hold. The assumption (12) is strong but natural for this type of estimate, and the result genuinely extends the classical LS inequality without being confused with the restricted fractional Laplacian versions in SV13 or MNS17.\n\nThe application half is where the paper gets shaky. Theorem 1.11 is only sketched, and Theorems 1.12 and 1.13 are stated without proof, leaning on AK19. That alone is a limitation, but the stress-test note found a concrete error: the functional in Lemma 5.1, equation (31), uses (1/2)∫|(-Δ)^s v|^2 dx. That is not the energy associated to A_σ = (-Δ)^s + 1/τ. Its Euler-Lagrange equation involves (-Δ)^{2s}, not the obstacle equation (30). For v ∈ H^s_0 the integral may not even be finite. The fix is straightforward—replace the integrand with |(-Δ)^{s/2}v|^2—and then the proof would follow from Theorem 1.6. But as written, the existence proof for the unidirectional diffusion equation is unsupported.\n\nSo the central theorem is good, and the paper deserves a serious referee. The refereeing should require a correction to (31) and a decision on how much of the AK19 arguments are imported. If the authors fix the functional and either expand the proofs or clearly state the applications as conditional on AK19, the LS inequality part is publishable. The self-citation to AK19 is legitimate—the extension is the point. The second half is currently a sketch with a real bug, so the well-posedness claim should be treated as unverified until the fix is in.\n\nTake it to review. A good referee can sort this out quickly.","headline":"The Lewy-Stampacchia inequality for the spectral fractional Laplacian is correct and new, but the application half has a concrete error in the discrete functional that makes the well-posedness proof unsupported as written.","tokens_in":23686,"tokens_out":5279,"would_cite":true,"duration_ms":160975,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R11","35K86","35K61"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves a Lewy-Stampacchia inequality for the spectral fractional Laplacian on bounded Lipschitz domains, giving two-sided pointwise bounds and L2 regularity for obstacle-problem solutions.","keywords":["Lewy-Stampacchia","obstacle problem","fractional Laplacian","spectral fractional Laplacian","variational inequality","unidirectional diffusion","well-posedness","Caffarelli-Silvestre extension"],"falsifier":"Solve the obstacle problem (14) numerically on a bounded Lipschitz domain with smooth forcing and smooth obstacle, and check pointwise whether $f \\le Au \\le \\max\\{f, A\\psi\\}$ holds; any single violation would disprove Theorem 1.6. Alternatively, choose an obstacle with $(-\\Delta)^s \\psi$ equal to a Dirac mass so that assumption (12) fails and test whether the inequality survives in any weaker sense, which would show whether the assumption is merely technical or truly necessary.","tokens_in":1860,"feed_emoji":"📐","tokens_out":7560,"duration_ms":101911,"temperature":0.7,"pith_summary":"This paper establishes a Lewy-Stampacchia inequality for the spectral fractional Laplacian on bounded Lipschitz domains. It shows that the solution of an obstacle variational inequality has its fractional Laplacian trapped between the forcing term and the obstacle's fractional Laplacian, pointwise almost everywhere. This extends the classical Laplacian result of Lewy and Stampacchia to the fractional setting, where the nonlocal nature of the operator creates genuine difficulties. The paper then applies the inequality to prove well-posedness, uniqueness, stability, comparison, and long-time convergence for anomalous unidirectional diffusion equations of fractional type. A sympathetic reader would care because the inequality is the key tool that upgrades distributional solutions to strong solutions with enough regularity to make the nonlinear evolution equations meaningful.","feed_headline":"Fractional obstacle solutions get a two-sided pointwise bound","feed_subtitle":"Spectral fractional Laplacian gets L2 regularity and well-posed anomalous diffusion.","key_machinery":"The argument is carried by the Caffarelli-Silvestre extension of the spectral fractional Laplacian, which realizes $(-\\Delta)^s$ as a Dirichlet-to-Neumann map of a degenerate elliptic problem in one extra dimension, together with Gustafsson's equivalent variational-inequality reformulation. Because the fractional Laplacian is nonlocal, the classical identity $\\langle -\\Delta u_+, u_-\\rangle = 0$ fails; Lemma 2.2 replaces it with $\\langle (-\\Delta)^s u_+, u_-\\rangle \\le 0$, proved through the extension. The equivalent constraint set $K_2 = \\{v : f \\le Av \\le \\max\\{f, A\\psi\\}\\}$ is then used to force the two-sided $L^2$ estimate and the membership $u \\in X^{2s}_0(\\Omega)$.","core_discovery":"Let $\\Omega$ be a bounded Lipschitz domain and $s \\in (0,1)$. For $f \\in L^2(\\Omega)$ and an obstacle $\\psi \\in H^s_0(\\Omega)$ whose fractional Laplacian is a signed Radon measure with positive part in $L^2(\\Omega)$, the unique solution $u$ of the obstacle variational inequality $\\langle Au, v-u\\rangle \\ge \\langle f, v-u\\rangle$ for all $v \\ge \\psi$ actually satisfies $u \\in X^{2s}_0(\\Omega)$ and, with $A = (-\\Delta)^s + \\lambda$, the two-sided bound $f \\le Au \\le \\max\\{f, A\\psi\\}$ almost everywhere in $\\Omega$. This is a complete Lewy-Stampacchia estimate for the spectral fractional Laplacian, including $L^2$ regularity of $(-\\Delta)^s u$. The same machinery yields comparison principles, uniqueness and stability, existence of strong solutions to the fractional unidirectional diffusion equation $\\partial_t u = [-(-\\Delta)^s u + f]_+$, and convergence of these solutions as $t \\to \\infty$ to an associated stationary obstacle problem.","pith_inferences":["A natural testable extension is whether the same two-sided estimate persists for more general nonlocal operators possessing a Caffarelli-Silvestre-type extension, such as stable-like operators with variable coefficients; the proof's reliance on the spectral representation suggests this may require new ideas.","In the limit $s \\to 1$, the fractional inequality should recover the classical Lewy-Stampacchia bound for the Laplacian, providing a consistency check for numerical discretizations of fractional obstacle problems.","The $L^2$ bound on $(-\\Delta)^s u$ implies additional spatial regularity that could be used to derive rates of convergence for finite element or finite difference methods for fractional obstacle problems, a consequence the paper does not develop."],"forward_implications":["For every obstacle and forcing satisfying the stated assumptions, the solution of the obstacle problem has $(-\\Delta)^s u \\in L^2(\\Omega)$, so expressions such as $[-(-\\Delta)^s u + f]_+$ are well defined pointwise almost everywhere.","The comparison principle Theorem 1.7 holds: larger forcing and larger obstacles give larger solutions, making the fractional obstacle problem order-preserving.","The anomalous unidirectional diffusion equation has a unique strong solution depending continuously on the data, by Theorems 1.10 and 1.11.","Solutions of the fractional unidirectional diffusion equation converge as $t \\to \\infty$ to the solution of a stationary obstacle problem, with the limit satisfying $u_\\infty \\ge u_0$ and $(-\\Delta)^s u_\\infty \\ge f_\\infty$.","The Lewy-Stampacchia estimate itself is exactly the two-sided bound $f \\le Au \\le \\max\\{f, A\\psi\\}$, the fractional analogue of the classical result."],"supporting_citations":[{"why":"Supplies the Caffarelli-Silvestre extension for the spectral fractional Laplacian and the norm equivalence used throughout the proof.","marker":"[CS16]"},{"why":"Provides the original extension problem that localizes the nonlocal fractional Laplacian.","marker":"[CS07]"},{"why":"Supplies the equivalent variational-inequality method used in Lemma 2.5 to derive the two-sided estimate.","marker":"[Gus86]"},{"why":"Gives the classical Laplacian case and the unidirectional diffusion framework that this paper extends to the fractional setting.","marker":"[AK19]"},{"why":"Original Lewy-Stampacchia inequality for second-order elliptic operators, the result being generalized.","marker":"[LS69]"},{"why":"Provides Lemma 2.2 and prior variational-inequality results for the spectral fractional Laplacian.","marker":"[MN17]"},{"why":"Develops the extension problem and trace results used to justify computations with the fractional Laplacian.","marker":"[ST10]"},{"why":"Establishes properties of the spectral fractional Laplacian that distinguish it from other fractional Laplacians.","marker":"[MN14]"},{"why":"Supplies Sobolev-space and extension-problem tools used in the definitions and regularity arguments.","marker":"[NOS15]"}],"fun_headline_variants":["Fractional obstacle solutions: two-sided pointwise bounds","Well-posed anomalous diffusion from fractional inequality","Nonlocal obstacle problem: bounds and regularity","Fractional Laplacian: pointwise estimate for obstacle"],"cache_read_input_tokens":25728,"weakest_assumption_plain":"The load-bearing premise is that the obstacle's fractional Laplacian is a signed Radon measure whose positive part is square-integrable; without it the upper bound $\\max\\{f, A\\psi\\}$ need not be an $L^2$ function and the proof's constraint set $K_2$ is undefined.","fun_headline_variants_meta":{"raw":{"variants":["Fractional obstacle solutions: two-sided pointwise bounds","Well-posed anomalous diffusion from fractional inequality","Nonlocal obstacle problem: bounds and regularity","Fractional Laplacian: pointwise estimate for obstacle"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000633,"raw_usage":{"total_tokens":2893,"prompt_tokens":886,"completion_tokens":2007,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":502,"completion_tokens_details":{"reasoning_tokens":1947}},"tokens_in":502,"tokens_out":2007,"duration_ms":14017,"temperature":1.0,"reasoning_tokens":1947,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:43:39.167559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the obstacle problem (14) numerically on a bounded Lipschitz domain with smooth forcing and smooth obstacle, and check pointwise whether $f \\le Au \\le \\max\\{f, A\\psi\\}$ holds; any single violation would disprove Theorem 1.6. Alternatively, choose an obstacle with $(-\\Delta)^s \\psi$ equal to a Dirac mass so that assumption (12) fails and test whether the inequality survives in any weaker sense, which would show whether the assumption is merely technical or truly necessary.","supporting_citations":[],"review_version":1}