{"id":"5740bfb9-638a-4e9d-973b-5ec4e807d8db","arxiv_id":"2605.00694","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Optimal controls in reaction-diffusion type problems are characteristic functions of sets with boundaries that are smooth up to a (d-2)-dimensional singular set.","lead":"The paper proves that optimal controls for a class of PDE-constrained maximization problems are bang-bang, meaning they equal the characteristic function of a set whose boundary is smooth except on a low-dimensional subset. This regularity result applies to problems arising in mathematical physics and biology and is obtained by reducing them to unstable free boundary problems.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest-assumption diagnosis matches the abstract's own description of the technical difficulty. Without the detailed proof steps, the only honest position is that the argument cannot yet be stress-tested for soundness; hence the verdict remains UNVERDICTED and no adjustment is warranted.","tokens_in":1855,"tokens_out":320,"duration_ms":34197,"concrete_test":"Extract the precise statement of the non-degeneracy lemma (presumably the analogue of Monneau–Weiss or Chanillo–Kenig–To) and test it on the simplest radial sign-changing example in 2D: solve the ODE for Θ(r) with right-hand side that changes sign once, compute the associated blow-up profile, and check whether the Weiss-type monotonicity formula still forces the density to be exactly 1/2 at the free boundary.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that a reduction to an unstable free-boundary problem yields bang-bang optimality and almost-everywhere smoothness of the free boundary, even though the functional is not minimizing and the Laplacian of the state is sign-changing. The abstract explicitly flags the non-degeneracy step as the novel technical point and states that it is obtained by blending optimal-control, free-boundary and measure-theoretic tools. Because the full manuscript text is supplied only as a placeholder and no explicit equations, lemmas or counter-examples are visible, no internal gap, unjustified assumption or inconsistency in the argument can be located.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript establishes the first general regularity results for a class of constrained optimal control problems of the form maximize ∫ ψ(Θ_m) − c ∫ m subject to −ΔΘ_m = m Θ_m + B(x, Θ_m) with 0 ≤ m ≤ 1 a.e. It proves that the optimal control m* is bang-bang (m* = χ_{E*}) and that the free boundary ∂E* is smooth outside a (d−2)-dimensional singular set. For the volume-constrained variant (fixed ∫ m = m_0), the same bang-bang property holds, and in two dimensions ∂E* consists of finitely many smooth curves. The argument proceeds by reduction to an unstable free-boundary problem, with a new non-degeneracy analysis for blow-ups that blends optimal-control, free-boundary, and measure-theoretic tools to handle the non-minimizing character of the functional and the sign-changing Laplacian.","tokens_in":1965,"tokens_out":557,"duration_ms":40093,"significance":"If the central claims hold, the paper supplies the first general existence-plus-regularity theory for these unstable free-boundary problems arising in mathematical physics and biology. The explicit handling of non-degeneracy in a non-minimizing, sign-changing setting extends the classical results of Monneau–Weiss and Chanillo–Kenig–To and demonstrates a workable synthesis of optimal-control and free-boundary techniques. This opens the door to further applications and to analogous regularity statements for related control problems.","major_comments":[],"minor_comments":[{"comment":"The abstract states that the non-degeneracy step is obtained by 'blending tools from optimal control theory, free boundary and measure theory,' but the main text should contain an explicit roadmap (e.g., a diagram or numbered list of lemmas) showing precisely which tool supplies which estimate; this would make the logical flow easier to follow.","section":null},{"comment":"In the volume-constrained case the two-dimensional regularity statement is stated only for d=2; a brief remark on the obstruction to extending the finite-union-of-curves conclusion to higher dimensions would be helpful.","section":null},{"comment":"The bibliography should include full citations for the cited works of Monneau–Weiss and Chanillo–Kenig–To, together with any recent extensions that are used in the blow-up analysis.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The manuscript appears to fit the scope of a journal in mathematical analysis; the citation pattern to the free-boundary literature is appropriate and the novelty claim is clearly delimited. No obvious overlap with existing results is visible from the abstract and outline."},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the careful and positive report, including the accurate summary of our results and the recognition of their significance in extending regularity theory to unstable free-boundary problems in optimal control. We appreciate the recommendation for minor revision.","responses":[],"tokens_in":1423,"tokens_out":65,"duration_ms":19411,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main point is that this work gives the first general regularity theorem for bang-bang optimal controls in a broad family of problems that show up in physics and biology. For the maximization problem with the state equation -ΔΘ_m = m Θ_m + B(x, Θ_m) and 0 ≤ m ≤ 1, the optimal m is the characteristic function of a set whose boundary is smooth outside a (d-2)-dimensional set. The volume-constrained version gets the same bang-bang property, and in two dimensions the free boundary is a finite union of smooth curves. They reach this by reducing the problem to an unstable free boundary problem and adapting the Monneau-Weiss and Chanillo-Kenig-To theory to the case where the functional is not minimizing and the Laplacian changes sign. The new technical step is their treatment of non-degeneracy for the blow-ups, which mixes optimal control, free boundary, and measure-theoretic tools. That step is what lets them close the argument in the unstable setting. The paper does a clean job of acknowledging the difficulties up front and organizing previous case-by-case results into one framework. The reduction itself looks straightforward once non-degeneracy is in hand, and the claims line up with what the abstract promises. The soft spot is the non-degeneracy argument itself; it carries a lot of weight, and any gap there would affect the whole regularity statement. The stress-test did not locate an internal contradiction or unjustified step, so the blending of tools appears to hold. This is specialized work aimed at people who already work on free boundary regularity or optimal control with PDE state constraints. A reader in that niche would find the unified treatment and the new non-degeneracy approach useful. It deserves a serious referee rather than a desk reject.","headline":"The paper delivers a first general regularity result for bang-bang solutions in this class of constrained optimal control problems by reducing them to unstable free boundary problems and supplying a new non-degeneracy argument.","tokens_in":2511,"tokens_out":443,"would_cite":false,"duration_ms":30978,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Optimal controls for a broad class of semilinear elliptic problems are bang-bang with free boundaries that are smooth outside sets of dimension at most d-2.","keywords":["optimal control","free boundary regularity","bang-bang property","semilinear elliptic equations","unstable free boundary problems","existence and regularity"],"falsifier":"An explicit example in which an optimal m* takes a value strictly between 0 and 1 on a positive-measure set, or in which the free boundary of E* contains a singularity of positive (d-1)-dimensional Hausdorff measure in dimension d > 2, would disprove the main claims.","tokens_in":2732,"feed_emoji":"📐","tokens_out":911,"duration_ms":47061,"temperature":0.7,"pith_summary":"The paper proves that maximizers of integrals of the form ∫ ψ(Θ_m) - c ∫ m, where Θ_m solves the semilinear equation -ΔΘ_m = m Θ_m + B(x, Θ_m) under the pointwise constraint 0 ≤ m ≤ 1, must be characteristic functions χ_{E*}. It further shows that the free boundary ∂E* is smooth except possibly on a closed set of Hausdorff dimension at most d-2. A parallel result holds for the volume-constrained version of the same problem, where in two dimensions the free boundary reduces to a finite union of smooth curves. The argument proceeds by recasting the optimality condition as an unstable free boundary problem and then establishing the required non-degeneracy and regularity properties for blow-ups. If these statements hold, many control problems in mathematical physics and biology reduce to the geometric task of selecting an optimal domain rather than an arbitrary density function.","feed_headline":"Bang-bang controls produce smooth free boundaries","feed_subtitle":"For maximization problems with semilinear state equations under pointwise bounds, optima are characteristic functions whose interfaces are C","key_machinery":"Reduction of the optimal-control problem to an unstable free-boundary problem, followed by blow-up analysis that yields non-degeneracy despite a sign-changing Laplacian and the absence of energy minimization.","core_discovery":"We prove that for a large class of problems of the form maximize ∫ ψ(Θ_m) - c ∫ m subject to -ΔΘ_m = m Θ_m + B(x, Θ_m) and 0 ≤ m ≤ 1 a.e., the solution m* equals χ_{E*} and ∂E* is smooth up to a (d-2)-dimensional subset. For the volume-constrained analogue maximize ∫ ψ(Θ_m) subject to the same PDE, the same pointwise bound on m, and fixed ∫ m = m_0, the optimal m* is again bang-bang and, when d = 2, ∂E* is a finite union of smooth curves. The proof is obtained by reducing the problem to an unstable free boundary problem whose regularity is analyzed via a new combination of optimal-control, free-boundary, and measure-theoretic tools that secure non-degeneracy of blow-ups even though the free ","pith_inferences":["The same reduction technique may apply to other control problems whose state equations are semilinear but whose optimality conditions produce non-variational free boundaries.","Numerical schemes could be designed to optimize directly over characteristic functions of domains whose boundaries satisfy the derived regularity.","The non-degeneracy arguments developed for sign-changing Laplacians could be tested on related free-boundary problems arising in phase-transition models.","If the singular set is empty for generic data, many optimal designs in applications would possess globally smooth interfaces."],"forward_implications":["The optimal control is necessarily 0 or 1 almost everywhere.","The free boundary is C^∞ outside a closed singular set of Hausdorff dimension at most d-2.","In the two-dimensional volume-constrained case the free boundary consists of finitely many C^∞ curves.","The same regularity conclusions apply to a large family of nonlinearities B and payoff functions ψ."],"fun_headline_variants":["Bang-bang controls smooth free boundaries","Smooth interfaces for bang-bang optima","Free boundaries smooth under control optima","Bang-bang property yields boundary regularity"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The reduction to the unstable free-boundary problem succeeds in proving non-degeneracy of blow-ups even though the free boundary is not minimizing and the Laplacian of the state function changes sign.","fun_headline_variants_meta":{"raw":{"variants":["Bang-bang controls smooth free boundaries","Smooth interfaces for bang-bang optima","Free boundaries smooth under control optima","Bang-bang property yields boundary regularity"]},"model":"grok-4.3","cost_usd":0.008642,"raw_usage":{"total_tokens":3913,"prompt_tokens":859,"num_sources_used":0,"completion_tokens":48,"cost_in_usd_ticks":86415500,"prompt_tokens_details":{"text_tokens":859,"audio_tokens":0,"image_tokens":0,"cached_tokens":64},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":3006,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":859,"tokens_out":48,"duration_ms":36412,"temperature":1.0,"reasoning_tokens":3006,"cache_read_input_tokens":64,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-05-09T18:51:21.024197+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"An explicit example in which an optimal m* takes a value strictly between 0 and 1 on a positive-measure set, or in which the free boundary of E* contains a singularity of positive (d-1)-dimensional Hausdorff measure in dimension d > 2, would disprove the main claims.","supporting_citations":[],"review_version":1}