{"id":"e0137541-533b-4855-bb33-948e08caa2e2","arxiv_id":"2607.20420","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Exact optimal control of two nonreciprocally coupled Brownian particles in moving harmonic traps is oscillatory, and unbounded work extraction appears beyond a critical time unless the trap range or active force is regularized.","lead":"Two microscopic particles held by movable traps can push and pull each other unevenly; this paper finds the exact minimum-work way to move them, and shows that beyond a certain duration the trap work can be made unboundedly negative. Realistic limits on the trap range turn that mathematical anomaly into sharp changes in the optimal strategy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unbounded-work anomaly is an artifact of the cost-free linear active force; whether it survives when the energy maintaining nonreciprocity is charged is deferred to SM and unverified.","rationale":"The reader's weakest assumption is exactly the load-bearing one. For T<Tc, the quadratic functional is convex and the exact solution is solid. For T>Tc, the negative mode makes the unregularized infimum −∞ mathematically; this is not in dispute. The issue is whether 'unbounded extractable work' is a physical thermodynamic result or an artifact of omitting the cost of the active interaction. The paper's only treatment is a deferred sentence to SM S7; without it, the anomaly is a property of a cost function, not the physical system. Therefore a conditional verdict is appropriate; no need to reject because the analytic solution and diagnostics stand if treated as model-level results. I agree with the reader's assessment and would not change the verdict.","tokens_in":9383,"tokens_out":20817,"duration_ms":201807,"concrete_test":"Augment the objective with a maintenance cost for the nonreciprocal interaction, e.g., W_μ = W + μ∫_0^T (αs)^2 dt (or, better, the housekeeping power derived from a microscopic reservoir model). Re-derive Eqs. (14)–(17) and compute the Hessian for T>Tc. If for any μ>0 all eigenvalues are positive (or the infimum becomes finite and the optimal amplitude scales as 1/μ), the unbounded anomaly is a singular limit of the μ=0 cost and is not a physical property of the dimer. Also rerun the zero-target bifurcation for small μ to see whether the first-/second-order transitions survive in the μ→0 limit.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (1) posits an unbounded linear active force αs with no associated energy cost, and the only nonreciprocal term in the cost (Eq. (7)) is −2α∫sẏdt. The T>Tc infimum −∞ follows from the negative mode of this quadratic form; that math is internally consistent. The load-bearing question is whether this constitutes a 'thermodynamic anomaly.' In a complete thermodynamic bookkeeping, the nonreciprocal interaction is maintained by a hidden reservoir whose power expenditure must appear in the objective. The paper's only response is one sentence—'Incorporating the energy input that maintains nonreciprocity into the objective function is another possible regularization'—and defers details to SM S7, which is not included. Since the headline result of unbounded extractable work and output power is qualified only by this absent calculation, the physical claim is unsubstantiated. If the energy cost is nonzero, the negative mode may be stabilized and the 'anomaly' disappears, reducing the central result to an artifact of a cost-free active-force model. This does not invalidate the T<Tc exact solution, but it does mean the title's 'Thermodynamic Anomaly' should be understood as a model-level singularity, not a physical prediction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper solves exactly a finite-time mean-work optimal control problem for two overdamped Brownian particles held by two independently movable harmonic traps, with either reciprocal or nonreciprocal pairwise interactions. The controls are eliminated in favor of the mean center and separation dynamics, yielding Euler-Lagrange equations, closed-form oscillatory optimal protocols, and an explicit quadratic work formula. For reciprocal interactions the solution reduces to known single-particle transport plus a finite-time separation cost. For nonreciprocal interactions, the central claim is that beyond a critical duration Tc the work functional acquires a negative mode, so the unregularized infimum of the external work is -infinity and extractable work/output power are unbounded. Finite trap range or force saturation regularize this divergence and, in the zero-target case, convert it into first-order-like or second-order-like optimal-protocol transitions, including multiple duration kinks. The paper also discusses experimental relevance to dual-trap actomyosin and levitated nanoparticle platforms.","tokens_in":9700,"tokens_out":11209,"duration_ms":89897,"significance":"If the central claims hold, this is a rare exactly solvable finite-time optimal-control problem for an interacting, genuinely nonequilibrium system. The closed-form solution, the clean reduction of the reciprocal case to prior single-particle results, and the explicit regularized bounds are strengths. The paper appears internally consistent in its variational derivation: the Euler-Lagrange equations, boundary conditions, and the determinant condition leading to Tc are mutually consistent, and there is no evident circularity or fitting. The main significance is therefore twofold: it provides a benchmark for interacting stochastic-thermodynamic control, and it identifies a sharp divergence and subsequent transitions that could, in principle, be probed in dual-trap or levitated-nanoparticle experiments. However, the headline 'thermodynamic anomaly' depends on treating the nonreciprocal force as cost-free; the physical interpretation is not yet established because the energy-input regularization is only mentioned and deferred.","major_comments":[{"comment":"The central unboundedness conclusion is asserted rather than proven in the reviewed text: the positive-definiteness of the second variation for all admissible perturbations, the divergence of A,B as (Tc-T)^-1, and the resulting global infimum -infinity for T>Tc are all assigned to SM Sec. S6, which is not present. Since claiming unbounded external work and output power is the paper's headline result, please provide these derivations in the main text or a complete appendix. A negative direction establishes local unboundedness along one mode, but the global variational statement, including endpoint-jump perturbations, requires the missing argument.","section":"Thermodynamic anomaly and its restoration, Eqs. (19)-(20), Fig. 3"},{"comment":"The anomaly is generated by Eq. (7)'s term -2 alpha integral s ydot dt, in which the nonreciprocal force alpha s is cost-free. The manuscript itself notes that incorporating the energy input that maintains nonreciprocity is a possible regularization, but defers this to SM Sec. S7. This is load-bearing for the title's 'Thermodynamic Anomaly': if that energy cost is nonzero, the negative mode may be stabilized and the unbounded-work divergence may disappear. The finite-range bound (20) and the saturating-force prescription are external constraints or model changes, not thermodynamic bookkeeping. Please either include an energy-accounting calculation or revise the claims to state explicitly that the anomaly is a property of the cost-free linear active-force model, not an established physical prediction.","section":"Thermodynamic anomaly and its restoration; Discussion"},{"comment":"The regularized-transition results -- hard-cutoff first-order-like jump, smooth-saturation scaling ||s*||_inf/s0 proportional to (T-Tc)^{1/2}, and the one-wall/two-wall switch producing a second kink -- are presented graphically or by assertion, with derivations deferred to SM Sec. S9, which is not included. Since these transitions are advertised as principal findings, the equations governing the regularized optima and the phase-boundary calculations should be shown, at least in an appendix, rather than only cited as SM.","section":"First- and second-order-like protocol transitions and multiple kinks, Fig. 4"}],"minor_comments":[{"comment":"The notation a,b and alpha,rho is mostly clear, but the stable regime rho=1+a+b>0 and the restriction rho>=1 used throughout should be stated more prominently; the reader must infer that the nonreciprocity parameters are dimensionless and that the extension-restoring condition is assumed from the start.","section":"Model, Eqs. (1)-(5)"},{"comment":"The Supplemental Material reference contains unresolved placeholders ('Refs. [? ? ?]'). This should be fixed, and the SM sections cited in the main text should be listed explicitly.","section":"References, Ref. [31]"},{"comment":"The statement that the inequality W_nr*(M=0) <= 2L^2/(T+2) strengthens monotonically with |alpha| is plausible but only sketched via the argument W0 - 2|alpha||integral s ydot dt|. A short derivation or a reference to the SM would improve readability.","section":"End Matter, Eq. (24)"}],"recommendation":"major_revision","confidential_remarks":"The variational core of the paper is internally consistent, and the exact T<Tc solution is a useful contribution. However, the manuscript relies heavily on SM sections S6, S7, and S9 for its headline claims, and the physical interpretation of the unbounded-work anomaly needs either an energy-accounting analysis or a clearly qualified claim that the anomaly is a model-level property of the cost-free active force. I recommend asking the editor to ensure the full supplemental material is available to reviewers; with the missing derivations and a suitably qualified interpretation, the paper could become acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuine new thing here is the exact finite-time optimal-control solution for two nonreciprocally interacting Brownian particles. That's not been done before for an interacting pair, and the derivation is straightforward to follow: eliminate controls in favor of mean paths, write the work functional, get linear Euler-Lagrange equations. The reciprocal limit reduces to known single-particle results (Eqs. 9-10), which is good independent support. The nonreciprocal solution — oscillatory interior protocols, asymmetric endpoint jumps, the critical-time determinant — follows from the algebra, and nothing looks fudged. The paper also does a nice job showing that protocol transitions appear after regularization, with the zero-target case as a clean test bed. No fitting, no self-citation, no circular reasoning as far as I can tell.\n\nThe soft spot is exactly what the stress-test note flags. The unbounded-work anomaly for T>Tc comes from the term -2α∫sγdot dt in the cost, with the active force αs unbounded and carrying no energy cost in the objective. That produces a negative mode in the quadratic form. The math is internally consistent, but calling it a thermodynamic anomaly is overreach unless you account for the hidden reservoir that maintains nonreciprocity. The paper's only response is one sentence deferring to SM Sec. S7, which is not included in the text I received. If that energy cost is nonzero, the negative mode may be stabilized and the anomaly could disappear — reducing the headline to an artifact of a cost-free linear active-force model. The T<Tc exact solution is not affected by this, but the title claim is.\n\nOther smaller issues: the transition results (first- and second-order-like, multiple kinks) are deferred to SM Secs. S6/S9, so I can't verify the numerics or the asymptotic scaling. The statement that results extend to higher dimensions and more particles is asserted without support. And the experimental connection to myosin platforms is speculative — fine as a discussion point, but not more.\n\nWho is this for? Researchers in stochastic thermodynamics and optimal control, especially those interested in interacting or active systems. The exact solution is worth having on record, and the anomaly question is worth settling with proper bookkeeping. With the SM included, this deserves serious refereeing; even without it, the core derivation is worth someone's time.\n\nRecommendation: send to peer review, but ask for the SM to be included and for the authors to either show the energy-cost regularization explicitly or soften the thermodynamic anomaly claim.","headline":"The exact nonreciprocal dimer control solution is a real, clean new result, but the unbounded-work anomaly is a model artifact unless the cost of maintaining nonreciprocity is accounted for — and that accounting is deferred to the missing SM.","tokens_in":10135,"tokens_out":1082,"would_cite":true,"duration_ms":12131,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Nonreciprocally interacting Brownian dimers have a critical transport duration: before it, an oscillatory protocol minimizes mean work; after it, the unregularized optimal work is unbounded below, with extractable work and power becoming in","keywords":["stochastic thermodynamics","optimal control","nonreciprocal interactions","Brownian dimers","work extraction","finite-time protocols","thermodynamic anomaly","optical tweezers"],"falsifier":"Numerically minimize the work functional W[y,s]=∫₀ᵀ(2ẏ²−2αsẏ+½ṡ²)dt with no bound on s and a purely linear force: for T just above the first positive zero of Δα(T,ρ), the minimum should go to −∞ as the endpoint jumps A and B are scaled up along the negative eigenmode, and finite-difference solutions should show no finite minimizer. If a finite minimum or bounded jumps are observed, the claimed anomaly would be contradicted.","tokens_in":9278,"feed_emoji":"⚡","tokens_out":5195,"duration_ms":46213,"temperature":0.7,"pith_summary":"This paper solves exactly the finite-time optimal-control problem for two overdamped Brownian particles held in independently movable harmonic traps and coupled by a nonreciprocal interaction. It shows that the optimal protocol is oscillatory and deliberately uses the extension channel even when only the center is to be moved, because the nonreciprocal interaction contributes an area term that can do work. The central claim is a thermodynamic anomaly: past a finite critical duration Tc, the mean-work functional is unbounded below, so both extractable work and output power are infinite unless the model is regularized. Adding a physical constraint—finite trap range or saturating active force—restores a finite optimum and turns the anomaly into well-defined transitions: a first-order-like jump for a hard range and a second-order-like continuous onset for smooth saturation. With a finite range, the optimal protocol can switch repeatedly as duration grows, producing multiple work-duration kinks.","feed_headline":"Past a critical time, extractable dimer work goes infinite","feed_subtitle":"Two nonreciprocally coupled traps can harvest unbounded power until finite range or force saturation restores a finite optimum.","key_machinery":"The central object is the mean-work functional W[y,s]=∫₀ᵀ(2ẏ²−2αsẏ+½ṡ²)dt plus boundary terms, with the nonreciprocal contribution −2α∫s ẏ dt coupling center and extension channels. This area term is what allows extension motion to help or oppose center transport. The Euler–Lagrange equations reduce to the oscillator v̈+α²v=0, and the stability of the quadratic form is governed by the determinant Δα(T,ρ)=2α²cos(αT)+α(ρ+1)sin(αT)+ρ[1−cos(αT)]; Tc is its first positive zero, the first conjugate time where the second variation loses positive definiteness.","core_discovery":"The paper derives a closed-form mean-work functional for the two-trap dimer and shows that for nonreciprocal coupling α≠0 the center velocity obeys a harmonic oscillator v̈ + α²v = 0, with optimal mean paths given by sinusoids. The minimal work is an explicit quadratic form in the prescribed center displacement L and excess-separation change M. The central discovery is that this quadratic form ceases to be positive definite at the first conjugate time Tc: for T<Tc the explicit sinusoid protocol is the unique global minimizer; at Tc the endpoint jumps diverge as (Tc−T)^{-1}; for T>Tc there is a negative direction, so the infimum of the external work is −∞. The paper then shows that imposing a","pith_inferences":["The divergence of the endpoint jumps as Tc is approached from below is best read as a diagnostic that the linear nonreciprocal force is a coarse-grained model; any real system will saturate or pay an energy cost, so the observable content of the anomaly is the regularized transition rather than the unregularized infinity.","The jump vector in the (Y, R/2) plane rotates by angle αT during the protocol, suggesting a geometric-phase interpretation of nonreciprocal optimal control and a possible link to area-pumping or topological effects in active-matter control.","The multiple-kink phenomenon should be generic whenever two competing strategies—area harvesting versus drag dissipation—are in balance under a control bound; analogous finite-time transitions may appear in discrete-state, quantum, or multi-particle optimal-control problems.","A dual-trap experiment with a nonreciprocally coupled pair, such as two optical traps around a motor-driven filament or optically coupled nanoparticles, could test for the predicted oscillatory extension excursions during pure center translation and for the sharp growth of control amplitudes near Tc."],"forward_implications":["For durations shorter than Tc, the explicit sinusoid protocol is the unique global work minimizer, so any alternative two-trap protocol costs more mean work.","For durations longer than Tc, the unregularized mean-work infimum is −∞, making the raw problem ill-posed and forcing any physically meaningful treatment to include a cutoff or saturation.","A hard finite trap range converts the anomaly into a first-order-like jump from the zero protocol to the maximum-range protocol in the zero-target case, while smooth force saturation gives a continuous, second-order-like onset with the order parameter growing as (T−Tc)^{1/2}.","Under finite-range constraints, successive optimal strategies (e.g., one-wall versus two-wall) can alternate as duration increases, producing multiple kinks in the work-duration curve that have no single-particle analog.","The long-time output power of the regularized optimal protocol approaches the plateau α²Rmax²/(2ρ²), with possible finite-time local shoulders or peaks.","Following the paper's claims, nonreciprocal control can extract finite negative work even before Tc, and the reciprocal limit recovers the familiar one-particle and free-particle results."],"fun_headline_variants":["Dimer work blows up past critical time, then transitions","Unbounded power from nonreciprocal Brownian dimer","Critical time makes dimer work infinite until saturation","Finite-time transitions in optimal dimer control","Nonreciprocal dimer: infinite work beyond critical time"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The anomaly rests on modeling the nonreciprocal interaction as an unbounded, cost-free linear force αs; if that force saturates or its maintenance energy is charged, the unbounded-work infimum disappears, as the paper's own regularized versions demonstrate.","fun_headline_variants_meta":{"raw":{"variants":["Dimer work blows up past critical time, then transitions","Unbounded power from nonreciprocal Brownian dimer","Critical time makes dimer work infinite until saturation","Finite-time transitions in optimal dimer control","Nonreciprocal dimer: infinite work beyond critical time"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000171,"raw_usage":{"total_tokens":1106,"prompt_tokens":738,"completion_tokens":368,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":482,"completion_tokens_details":{"reasoning_tokens":294}},"tokens_in":482,"tokens_out":368,"duration_ms":3679,"temperature":1.0,"reasoning_tokens":294,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T09:51:59.429926+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically minimize the work functional W[y,s]=∫₀ᵀ(2ẏ²−2αsẏ+½ṡ²)dt with no bound on s and a purely linear force: for T just above the first positive zero of Δα(T,ρ), the minimum should go to −∞ as the endpoint jumps A and B are scaled up along the negative eigenmode, and finite-difference solutions should show no finite minimizer. If a finite minimum or bounded jumps are observed, the claimed anomaly would be contradicted.","supporting_citations":[],"review_version":1}