{"id":"f985c65f-189f-4a13-abbe-49364720582a","arxiv_id":"2508.13441","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A stochastic homogenization theorem and well-posedness result for a one-dimensional Hele-Shaw problem with random oscillations in both the interior and on the free boundary.","lead":"This paper proves that a one-dimensional Hele-Shaw free boundary problem with oscillating random coefficients has unique solutions and that, at small scales, the random oscillations average out to a deterministic effective equation. The result extends known homogenization theory to problems where both the interior diffusion and the moving boundary are heterogeneous.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Assumption 1.5(iv) is load-bearing: without G strictly positive or identically zero, Lemma 5.7's comparison condition (1.6) for the homogenized velocity is unproved, so Theorem 1.6's uniqueness step is conditional.","rationale":"I read the paper in good faith and followed the main line of the proof. The central claim is Theorem 1.6, and its proof relies on Proposition 6.1, which shows the half-relaxed limits of the supports are sub- and superflows of the effective problem. To conclude that these two flows coincide, the proof applies the comparison principle (Theorem 3.3) for the limiting equation, which requires the homogenized velocity V to satisfy the structural condition (1.6). Lemma 5.7 proves this condition, but only under Assumption 1.5(iv). The gap identified by the reader is exactly the one the authors flag as technical in Remark 1.7(i): no proof covers G that is nonnegative but neither strictly positive nor identically zero. This is load-bearing because without (1.6) the uniqueness of the limiting flow is not established, so the convergence of p_epsilon to a single deterministic p is not justified. The reader's other note about Lemma 4.4 is, in my reading, a minor issue that can be repaired by the covering argument already implicit in the proof; it is not the main obstacle. The missing +y in the unnumbered shift relation after (5.3) also appears to be a typographical slip that does not alter the subsequent estimates. Thus the single most serious concern is the unproved structural condition on the effective velocity under the intermediate case of G. I agree with the reader's CONDITIONAL verdict and recommend no change; the paper should be accepted only if the technical assumption is either removed, relaxed, or explicitly shown to be necessary.","tokens_in":41300,"tokens_out":20525,"duration_ms":197064,"concrete_test":"Test whether the excluded case actually breaks the comparison condition. Take deterministic coefficients a=b=1 and a 1-periodic nonnegative g with g=0 on a set of positive measure and g>0 on its complement, e.g., g(x)=1_{[0,1/2]} extended periodically. Compute V(q)=1/(integral_0^1 dx/(q+g(x))) and check Assumption 1.1(ii)(1.6), i.e., whether (1+gamma)^2 V(q) + C*gamma^{-1}|x1-x2|^2 >= V((1+gamma)q) holds for all q>0 and gamma in (0,1). If it holds for this and similar examples, then Assumption 1.5(iv) is likely removable and Lemma 5.7 needs an alternative proof; if it fails for some such g, then the restriction is essential and Theorem 1.6's uniqueness assertion is not valid in the intermediate case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 1.6) requires the homogenized problem (1.3) to have a unique viscosity solution, so that the half-relaxed limits of p_epsilon are forced to coincide. Uniqueness rests on Assumption 1.1(ii), specifically the comparison condition (1.6), which Lemma 5.7 establishes for the effective velocity V only under Assumption 1.5(iv): either G is uniformly strictly positive or G identically zero. In the proof of Lemma 5.7, strict positivity is used in case (2b) to turn an additive error C*gamma into a multiplicative factor (1+C*gamma)(q1*f+g); if g vanishes on a set of positive measure, this step fails and the estimate |V^{(1)}_{q1} - V^{(2)}_{q2}| <= C*delta*(1+min{|q1|,|q2|}) + C*gamma, and hence (1.6), is not derived. The authors acknowledge this in Remark 1.7(i) as technical, but the gap is real: no argument is given for the excluded intermediate case G>=0, not identically zero, not strictly positive. Since the comparison principle for the effective problem is used in the final step of the proof of Theorem 1.6 to identify the limit, the homogenization theorem as stated does not cover that intermediate case. This is a proof gap in the central argument, not merely a stylistic restriction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a one-dimensional Hele-Shaw type free boundary problem in which the elliptic operator in the positivity set and the free boundary velocity both have rapidly oscillating, stationary ergodic coefficients. The authors introduce a new notion of viscosity flows for the positivity set, prove a comparison principle and well-posedness for these flows, and then establish a stochastic homogenization theorem: solutions of the microscopic problem converge locally uniformly for almost every time to the unique viscosity solution of a deterministic homogenized problem, with an effective velocity obtained from an auxiliary frozen-coefficient ODE via the subadditive ergodic theorem. The homogenization theorem is proved under Assumption 1.5, which includes the dichotomy that G is either uniformly strictly positive or identically zero.","tokens_in":41538,"tokens_out":32778,"duration_ms":314163,"significance":"If the proofs are correct, this is a substantial contribution: it provides a viscosity-flow framework for well-posedness of inhomogeneous Hele-Shaw problems, and it gives the first homogenization result that treats simultaneous interior and free-boundary oscillations in a random medium. The result is new even in the periodic setting. A particular strength is that the effective velocity is not fitted or guessed: it is derived from a frozen-coefficient ODE and the subadditive ergodic theorem, and the paper contains detailed proofs of the principal estimates. The main limitation is the dependence on Assumption 1.5(iv): homogenization is established only when G is uniformly positive or identically zero, and the intermediate case is explicitly left open. The abstract and introduction should state this dichotomy clearly, since the phrase 'coefficients are stationary ergodic' alone suggests a broader result.","major_comments":[{"comment":"The density argument used to pass from Wiener's ergodic theorem to a uniform good set is not valid as written. In the proof of Lemma 4.4, inequality (4.3) bounds the bad set in B_{2MR} by at most 8δMR, but this does not imply that for every x∈B_{MR} there is a good z with |z-x|≤4δMR: a bad interval of length 8δMR centered at x is compatible with (4.3) and contains no good point within distance 4δMR. The same flawed step appears in the proof of Proposition 5.6 around (5.8). Since both results are used in Lemma 6.3, the homogenization proof is incomplete at a load-bearing point. The gap is readily repairable (for example, by taking the Egorov good set to have measure at least 1-δ/2, or by replacing 4δ with 8δ and adjusting the final constants), but the current text needs a corrected argument.","section":"Lemma 4.4 and Proposition 5.6"}],"minor_comments":[{"comment":"As stated, Assumption 1.1(ii) requires V(x1,q1)≥η(|q1|) for all q1∈R, but for the microscopic problem with G≡0 one has V(x,0)=0. The assumption should either exclude q1=0 or allow η(0)=0, since the comparison argument only uses this inequality for nonzero slopes.","section":"Assumption 1.1(ii) and Lemma 3.5"},{"comment":"The paper should prominently disclose that Theorem 1.6 is conditional on Assumption 1.5(iv). The abstract and introduction presently refer only to stationary ergodicity, which is broader than the actual hypotheses; the excluded intermediate case G≥0, with G neither strictly positive nor identically zero, is not covered and no argument is supplied for the authors' belief that the assumption is technical.","section":"Remark 1.7(i) and abstract"},{"comment":"The proof of Lemma 4.6 is somewhat handwavy when claiming that the liminf is determined by its values on rational points and can be represented as an infimum of subsequential limits. Since the solutions are uniformly Lipschitz in space with a uniform constant, the standard Arzelà-Ascoli argument should be stated more explicitly.","section":"Lemma 4.6"},{"comment":"There are several typos: 'hueristics' in Section 1.2.2 and Section 5.1, 'subaddive' in Section 1.2.2, and the expression 'S x0 q,ε/t' in the proof of Proposition 5.6 should be written more clearly (it denotes S_{q,ε'}^{x0}(1) with ε'=ε/t).","section":"Typos"}],"recommendation":"major_revision","confidential_remarks":"I found no fatal flaw in the central strategy, and the main ideas are original and significant. The proof of Lemma 4.4 and Proposition 5.6 needs correction, but the repair is straightforward. The dependence on Assumption 1.5(iv) should be more visible in the abstract and introduction. The paper is otherwise carefully written and deserves publication after these revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my take. This is a substantial paper: it proves well-posedness and stochastic homogenization for a 1D Hele-Shaw problem where the interior elliptic operator and the free boundary velocity both oscillate. The results are new even in the periodic case, and the technical novelty is real — the viscosity-flow formulation for space-time sets gets around the failure of sup-inf convolutions for inhomogeneous interior operators, and the subadditive arrival-time quantity is a clever way to couple the two homogenization mechanisms.\n\nThe proofs are long but mostly careful; I did not find a fatal error. The literature review is accurate, and the authors are honest about prior work.\n\nThe main soft spot is the assumption on G. The stress-test note is correct: in Lemma 5.7, strict positivity of G is used in case (2b) to turn an additive error into a multiplicative factor. If G is merely nonnegative, not identically zero, the estimate (5.9) is not proved, and the comparison condition (1.6) for the effective velocity is not established. Since uniqueness of the homogenized problem is used in the final step of Theorem 1.6, the theorem does not cover that intermediate case as written. The authors flag this as technical in Remark 1.7(i), but that undersells it: without an additional argument, the gap is real. It is not a fatal flaw — the theorem is stated under Assumption 1.5(iv) — but a referee should ask for a proof of the intermediate case or an explicit statement that it is open.\n\nThe minor issue in Lemma 4.4 is also real but genuinely minor: the density bound gives a bad set of size at most 8δMR, so to guarantee a good point within distance of every x, one needs a radius of 8δMR, not 4δMR. That is a constant fix.\n\nWho this is for: specialists in free boundary problems and stochastic homogenization. The viscosity-flow framework may have legs beyond this paper. It deserves a serious referee, and I would accept it for review with the expectation of a revision addressing the G assumption and the Lemma 4.4 constant.","headline":"Valuable paper with real novelty, but the homogenization theorem's uniqueness step hinges on a positivity assumption on G that the authors understate.","tokens_in":42126,"tokens_out":3698,"would_cite":true,"duration_ms":36613,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R35","35B27","35D40","60F15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Random oscillations in both the interior operator and the free boundary velocity of a Hele-Shaw flow self-average, so that almost surely and for almost every time the random solutions converge to one deterministic free boundary evolution.","keywords":["Hele-Shaw flow","free boundary problem","viscosity solutions","viscosity flows","stochastic homogenization","stationary ergodic coefficients","subadditive ergodic theorem","one-dimensional PDE"],"falsifier":"Take a stationary ergodic medium in which $G$ is nonnegative, not identically zero, and not uniformly positive (for example the simplest one-dimensional arrangement where $G(x,y,\\omega)=h(y)$ with $h\\ge 0$ equal to zero on a Cantor set and positive elsewhere), with $A,B$ positive; compute the effective velocity $\\bar V$ from the ODE (5.1) and check whether it satisfies condition (1.6). If the condition fails, the limiting problem may admit more than one viscosity flow, and the convergence in Theorem 1.6 would fail as stated; if it still holds, the strict-positivity assumption is removable.","tokens_in":41031,"feed_emoji":"🌀","tokens_out":7702,"duration_ms":77118,"temperature":0.7,"pith_summary":"The paper studies a one-dimensional Hele-Shaw flow — the classical model of a viscous fluid squeezed between two narrowly separated plates, also used for tumor growth — in a medium whose coefficients oscillate rapidly and randomly. The first contribution is a well-posedness theorem: for bounded open initial support, a unique 'viscosity flow' exists, defined on the positivity set rather than on the solution function. The second, main contribution is a stochastic homogenization theorem: as the oscillation scale tends to zero, solutions converge locally uniformly, for almost every time and almost every realization of the random medium, to the unique solution of a deterministic free boundary problem with averaged coefficients. The paper thus establishes a rigorous sense in which small-scale randomness in both the interior elliptic operator and the boundary speed self-averages into one effective front law.","feed_headline":"Random Hele-Shaw fronts homogenize almost surely","feed_subtitle":"Oscillation in both the interior operator and boundary speed averages to one deterministic flow.","key_machinery":"Two mechanisms carry the argument. (i) Viscosity flows: instead of testing the solution function itself, the authors assign to each space-time set $\\Omega(t)$ the exact solution of the elliptic equation in $\\Omega(t)$ with zero boundary data, and define sub- and super-flows via the free boundary inequality against smooth test functions. This lets comparison proceed at the level of domains and avoids the fact that sup- and inf-convolutions destroy sub- and supersolution structure for inhomogeneous operators. (ii) The effective velocity via subadditive arrival times: with coefficients frozen at a point, one solves the ODE $\\frac{dS}{dt} = q\\, b(S,\\omega)\\,\\bar a/a(S,\\omega) + g(S,\\omega)$; the arrival time to a point $x$ is subadditive under spatial shifts, so the subadditive ergodic theorem gives an almost-sure linear growth speed $\\bar V(x_0,q)$. A uniform-in-space version of this convergence and a uniform pointwise ergodic theorem for the interior averages are what allow the half-relaxed limits of $p_\\varepsilon$ to be recognized as flows of the effective problem.","core_discovery":"The central claim, Theorem 1.6, is that for any bounded open initial set $O$ there is a full-measure set of environments for which $p_\\varepsilon(\\cdot,t,\\omega)\\to p(\\cdot,t)$ locally uniformly for almost every $t>0$. Here $p_\\varepsilon$ solves the problem with oscillating coefficients $A(x,x/\\varepsilon,\\omega)$, $B(x,x/\\varepsilon,\\omega)$, $F(x,x/\\varepsilon,\\omega)$, and $G(x,x/\\varepsilon,\\omega)$, and $p$ solves the homogenized problem with deterministic coefficients $\\bar A$, $\\bar F$, and $\\bar V$. The interior coefficients are explicit: $1/\\bar A(x)=\\mathbb{E}[1/A(x,\\cdot,\\cdot)]$ and $\\bar F(x)=\\mathbb{E}[F(x,\\cdot,\\cdot)]$. The effective free boundary velocity $\\bar V(x,q)$ is not a simple average: it is the almost-sure speed of a translated ODE whose coefficients are the frozen random data, obtained by applying the subadditive ergodic theorem to arrival times. The proof works by showing that the limsup and liminf limits of the supports are, respectively, a viscosity subflow and superflow of the effective problem, and then invoking the comparison principle for flows.","pith_inferences":["If the strict-positivity assumption on $G$ is truly technical, the proof strategy should extend to nonnegative $G$ with isolated zeros as soon as a continuity estimate for $\\bar V$ is available without the dichotomy; the paper does not supply that estimate.","The homogenization and the incompressible tumor-pressure limit are both singular limits; the paper leaves open whether $\\varepsilon\\to 0$ and $k\\to\\infty$ commute, and the present result gives a plausible starting point for testing commutation numerically.","The explicit ODE representation of the effective speed is specific to one dimension; a multidimensional analogue would need a different averaging object, likely one tied to normals and local geometry rather than a single arrival time."],"forward_implications":["For almost every environment and almost every time, the random oscillating problem has a sharp deterministic large-scale description; individual realizations differ only by errors that vanish locally uniformly.","The well-posedness theorem provides a unique viscosity flow for the heterogeneous problem, giving a predictive framework for fronts in nonuniform media beyond the averaging limit.","The periodic version of the homogenization theorem follows as a special case; the authors state that this was previously unknown when both interior and boundary oscillate.","The explicit formulas for $\\bar A$ and $\\bar F$ mean the effective interior behavior can be computed directly from the law of the medium, while the velocity must be computed from the ODE speed.","Because the support expands immediately and uniformly in $\\varepsilon$ and $\\omega$, there is no initial-time boundary layer in the support for the convergence result."],"supporting_citations":[{"why":"introduces the viscosity-solution framework and comparison strategy for the constant-coefficient Hele-Shaw problem that this paper generalizes.","marker":"[22]"},{"why":"is the closest previous stochastic homogenization result for Hele-Shaw flow, with no interior oscillation, and the obstacle-problem route that is unavailable with interior heterogeneity.","marker":"[29]"},{"why":"provides the subadditive ergodic theorem used to extract the almost-sure effective arrival time and hence the effective velocity.","marker":"[30]"},{"why":"supplies the pointwise ergodic theorem for continuous parameters used to obtain the averaged interior coefficients and their uniform convergence.","marker":"[5]"},{"why":"supplies the pointwise ergodic theorem in the probability-textbook form used in Lemma 4.4 on uniform convergence to the mean.","marker":"[21]"},{"why":"treats periodic homogenization of the free boundary velocity and marks the compactness contrast with the random setting.","marker":"[24]"}],"fun_headline_variants":["Oscillating Hele-Shaw fronts homogenize almost surely","Almost-sure homogenization for oscillating Hele-Shaw","Hele-Shaw fronts with random oscillations homogenize","Oscillating interior and boundary data homogenize Hele-Shaw"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The thesis rests on the dichotomy that the boundary term $G$ is either uniformly strictly positive or identically zero; if $G$ vanishes on some places but not all, the proof does not cover the uniqueness of the effective problem, and the homogenization theorem is not established.","fun_headline_variants_meta":{"raw":{"variants":["Oscillating Hele-Shaw fronts homogenize almost surely","Almost-sure homogenization for oscillating Hele-Shaw","Hele-Shaw fronts with random oscillations homogenize","Oscillating interior and boundary data homogenize Hele-Shaw"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000955,"raw_usage":{"total_tokens":4049,"prompt_tokens":898,"completion_tokens":3151,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":514,"completion_tokens_details":{"reasoning_tokens":3079}},"tokens_in":514,"tokens_out":3151,"duration_ms":25863,"temperature":1.0,"reasoning_tokens":3079,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:14:14.475417+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a stationary ergodic medium in which $G$ is nonnegative, not identically zero, and not uniformly positive (for example the simplest one-dimensional arrangement where $G(x,y,\\omega)=h(y)$ with $h\\ge 0$ equal to zero on a Cantor set and positive elsewhere), with $A,B$ positive; compute the effective velocity $\\bar V$ from the ODE (5.1) and check whether it satisfies condition (1.6). If the condition fails, the limiting problem may admit more than one viscosity flow, and the convergence in Theorem 1.6 would fail as stated; if it still holds, the strict-positivity assumption is removable.","supporting_citations":[{"cited_title":"Kim , Uniqueness and existence results on the hele-shaw and the stefan problems , Archive for Rational Mechanics & Analysis, 168 (2003)","cited_arxiv_id":null,"evidence_quote":"introduces the viscosity-solution framework and comparison strategy for the constant-coefficient Hele-Shaw problem that this paper generalizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"is the closest previous stochastic homogenization result for Hele-Shaw flow, with no interior oscillation, and the obstacle-problem route that is unavailable with interior heterogeneity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the subadditive ergodic theorem used to extract the almost-sure effective arrival time and hence the effective velocity."},{"cited_title":"Bergelson, A","cited_arxiv_id":null,"evidence_quote":"supplies the pointwise ergodic theorem for continuous parameters used to obtain the averaged interior coefficients and their uniform convergence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"treats periodic homogenization of the free boundary velocity and marks the compactness contrast with the random setting."}],"review_version":2}