Pith. sign in

REVIEW 3 major objections 5 minor 32 references

Optimal Trapping of Brownian Motion: A Nonlinear Analogue of the Torsion Function

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The ball, with drift pointing radially inward, is the optimal trap for Brownian motion with a drift of fixed strength.

desk verdict A new and likely true isoperimetric theorem for a nonlinear elliptic PDE, but the written proof has a fixable gap in the reduction from arbitrary drifts to the nonlinear equation and a hand-waved differentiation step. read the letter →

arxiv 1908.06273 v1 pith:VNZA6TEC submitted 2019-08-17 math.AP math.PR

classification math.APmath.PR MSC 35B5149K2060J60
keywords driftdiffusionexittimeisoperimetricinequalitytorsionfunctionBrownianmotionexpectedlifetimenonlinearellipticPDElevel-setmethod
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks how to keep a Brownian particle inside a bounded container for as long as possible when an external flow of prescribed maximum strength can push it around. The answer it proposes is sharp: among all containers of a given volume and all flows of a given maximum speed, the expected survival time is largest when the container is a ball and the flow always points straight toward the center of the ball. In any container, the best flow is not chosen in advance but is nonlinearly tied to the survival-time function itself, pointing along its gradient; this reduces the problem to a single nonlinear PDE. The paper proves that for this PDE the ball maximizes the total variation of the survival-time function and the integral of its Laplacian, and it recovers the known fact that the ball maximizes all $L^p$ norms of the survival time.

What carries the argument

The central object is the nonlinear PDE $-\Delta u - b|\nabla u| = 1$ with $b = \|b\|_{L^\infty}$ constant, called a nonlinear analogue of the torsion function (the torsion function solves $-\Delta u = 1$). Its defining property for the proof is invariance under adding constants: if $u$ solves it on $\Omega$, then $u - \varepsilon$ solves the same PDE on the superlevel set $\Omega_\varepsilon = \{u \geq \varepsilon\}$. This lets the authors induct over level sets, trading the shape optimization for a differential inequality $f'(c) \leq b(f(c)+c)/(c_d c^{(d-1)/d})$ for $f(c)$, the supremum of $\int_\Omega |\nabla u|\, dx$ over domains of volume $c$. Two classical ingredients make the induction valid: the Zaremba-Hopf-Oleinik boundary point lemma, which ensures the inward normal derivative of $u$ does not vanish, and $C^2$ regularity of solutions, which makes the volume expansion $|\Omega_\varepsilon| = |\Omega| - \varepsilon \int_{\partial\Omega} (\partial u/\partial n)^{-1} d\sigma + o(\varepsilon)$ exact enough. The ball is the extremal case because its radial solution satisfies the same ODE with equality at every step.

What would settle it

Compute, for a fixed volume and fixed maximal drift strength, the solution of $-\Delta u - b|\nabla u| = 1$ on a non-ball domain such as a long thin rectangle or an ellipse with the same area as a reference ball, and compare $\int_\Omega |\nabla u|\, dx$ or $\int_\Omega -\Delta u\, dx$ with the ball's value; if either integral exceeds the ball's value, the theorem is false. A numerical or explicit counterexample would settle it.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes the following theorem: for bounded $C^2$ domains with fixed volume and vector fields with fixed $L^\infty$ norm, the solution $u$ of $-\Delta u + b\cdot\nabla u = 1$ with $u=0$ on the boundary maximizes both $\int_\Omega |\nabla u|\, dx$ and $\int_\Omega -\Delta u\, dx$ when $\Omega$ is the ball and $b = -\|b\|_{L^\infty} \nabla u/|\nabla u|$. The optimal flow condition is derived from a maximum-principle comparison: replacing any candidate $b$ by the feedback field aligned with $\nabla u$ can only increase $u$ pointwise. This reduces the shape problem to the nonlinear PDE $-\Delta u - b|\nabla u| = 1$, whose invariance under adding constants lets the proof peel off level sets and obtain a differential inequality for the maximal gradient integral; the ball saturates the inequality via an explicit radial ODE. A corollary, already known from a general rearrangement principle, is that the $L^p$ norms of $u$ are also maximized by the ball.

Load-bearing premise

The load-bearing premise is that on every candidate domain the survival-time solution is smooth up to the boundary and meets it with nonzero inward derivative, so the level-set layers have the precise volume the proof needs; this is proved for $C^2$ domains, and the true minimal regularity is open.

Editorial extensions

If this is right

  • In any fixed domain, replacing the drift by $b = -\|b\|_{L^\infty} \nabla u/|\nabla u|$ increases the expected lifetime pointwise, so optimal traps can always be taken to push in the direction of increasing lifetime.
  • Among equal-volume domains, the ball maximizes both $\int_\Omega |\nabla u|\, dx$ and $\int_\Omega -\Delta u\, dx$, giving sharp bounds on the total variation of expected lifetime.
  • The same level-set argument shows that the ball maximizes $\|u\|_{L^p}$ for every positive integer $p$, with an elementary derivation for the $L^\infty$ case.
  • On the ball the optimal configuration is explicit: a radial solution of a one-dimensional ODE gives the exact maximal values as functions of volume and drift strength.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The authors leave the optimal boundary-regularity threshold open; a natural test is whether the volume expansion and differential inequality persist on $C^1$ or Lipschitz domains, where corners might change trapping efficiency.
  • Because the proof uses only level-set invariance and the boundary-point lemma, it may extend to other functionals of $u$ that are additive over level sets, such as moments of the exit time or the work done by the drift.
  • A concrete numerical probe of the isoperimetric claim would be to solve the nonlinear PDE on ellipses of the same area as a ball and check that $\int_\Omega |\nabla u|\, dx$ stays below the ball's value, which would also show how the optimum is approached as the domain becomes round.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies the Dirichlet problem -Δu + b·∇u = 1 in a bounded C^2 domain Ω, which models the expected exit time of drift-diffusion. For fixed ||b||_∞ and fixed volume, the authors claim that ∫_Ω |∇u| and ∫_Ω -Δu are maximized when Ω is a ball and b = -||b||_∞ ∇u/|∇u|. The proof introduces the nonlinear PDE -Δu - b|∇u| = 1 (with b now a constant equal to the L∞ norm of the drift), defines f(c) as the supremum of ∫|∇u| over domains of volume c, derives a differential inequality for f(c), and shows that the ball saturates it. The argument is extended to L^p norms; the L^p result is credited to Hamel and Russ. The paper is self-contained apart from the gaps noted below, and the ball computation is explicit and parameter-free.

Significance. If the main theorem is fully established, it provides a sharp isoperimetric statement for a natural nonlinear analogue of the torsion function and an explicit optimal drift design for maximizing expected exit times. The level-set induction is elegant, and the explicit ball computation in Lemma 4 is a clear strength. The L^p corollary is already known from Hamel–Russ, so the genuinely new content is the L^1 gradient maximization. The proof is structurally plausible, but the reduction from arbitrary drifts to the nonlinear PDE is currently missing, and the monotonicity issue in Lemma 3 needs a fix; both are load-bearing for the central claim.

major comments (3)
  1. [3.1, Lemma 1 and the paragraph after it] Lemma 1 does not establish the stated reduction to the nonlinear PDE. For a solution w of -Δw + b·∇w = 1, the lemma constructs u solving -Δu - ||b||_∞ (∇w/|∇w|)·∇u = 1 and shows u ≥ w; the drift is frozen at ∇w/|∇w|, so u is not a solution of -Δu - ||b||_∞ |∇u| = 1. The manuscript asserts without proof that this 'reduces the problem' to the nonlinear PDE. To justify the quantification over all b in the main theorem, the authors must either prove that the self-consistent solution of the nonlinear PDE dominates every linear solution, or prove convergence of the iteration w ↦ u to a solution of the nonlinear PDE. A direct maximum-principle comparison between the nonlinear solution and w would supply pointwise domination, but it is not included, and pointwise domination alone would still need to be supplemented to control ∫|∇u|. As written, Lemmas 2–6 apply only to nonlinear solutions, so the claim for arbitrary b is unsupported.
  2. [3.2, proof of Lemma 3] The proof of the differential inequality for f uses the step f(|Ω_ε|) ≤ f(c - ε|∂Ω|^2/(f(c)+c) + o(ε)), which requires f to be nondecreasing in c. The text only cites continuity of f at this point; continuity is insufficient to pass from |Ω_ε| ≤ c - εA + o(ε) to the displayed inequality with f evaluated at c - εA. The authors should either prove monotonicity of f (for example by a domain-inclusion comparison for the nonlinear PDE) or replace this step with a limsup argument that avoids monotonicity. This gap affects the derivation of the main estimate for ∫|∇u|.
  3. [2.1 and 3.2 (existence and regularity)] The manuscript never states an existence theorem for the nonlinear PDE -Δu - b|∇u| = 1 on an arbitrary bounded C^2 domain, and the C^2 regularity up to the boundary used in the volume expansion of Lemma 3 is asserted with a reference ([13]) to fully nonlinear theory that does not directly apply to this semilinear equation. Since Lemma 3's expansion |Ω_ε| = |Ω| - ε∫∂Ω (∂u/∂n)^{-1} dσ + o(ε) relies on this regularity and on ∂u/∂n > 0, the authors should provide a precise existence/regularity statement (e.g., via sub/supersolutions and elliptic regularity) or a specific reference. The issue is fixable by standard methods, but it is currently load-bearing because the theorem ranges over all C^2 domains.
minor comments (5)
  1. [Abstract] The formula '−∆u − b · |∇u| = 1' uses a dot product with a scalar b; once b denotes the L∞ norm, write 'b |∇u|' without the dot.
  2. [Introduction] There are several typos and OCR artifacts in the text, such as 'maximal strenght' and 'nc an'; a careful proofreading pass is needed.
  3. [3.2, Lemma 4 proof] The change of variables ∂/∂c = (1/|∂B_R|) ∂/∂r is written informally; a sentence justifying it via the coarea formula or the implicit function theorem would improve clarity.
  4. [References] Reference [13] is not the standard source for C^{2,α} regularity of semilinear elliptic equations with Lipschitz gradient dependence; a reference such as Gilbarg–Trudinger would be more appropriate.
  5. [2.2] The paragraph on existing results is compressed; a brief statement of the Hamel–Russ rearrangement result and its hypotheses would help the reader see exactly why the L^p corollary follows.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the proof is self-contained, the ball computation is explicit, and the only self-citation is background.

full rationale

The claimed derivation chain is not circular. Lemma 1 is a pointwise comparison proved by the maximum principle: given a linear solution w, the solution u for the drift -||b||_infty grad w/|grad w| satisfies u >= w. This is an independent monotonicity statement, not a definition of the nonlinear solution. The subsequent assertion that Lemma 1 'reduces the problem' may be incomplete—the drift is frozen at grad w rather than at grad u, so self-consistency is not fully proved—but that is a correctness gap, not a circular reduction. The paper does not fit a parameter to data and then rename it a prediction; no fitted constants enter. Lemmas 3-6 bound universal suprema f, g, h_p by differential inequalities derived from the PDE, coarea, and isoperimetry; Lemma 4 is an explicit radial computation showing that the ball saturates the same ODE. The Lp corollary is explicitly credited to Hamel and Russ, so it is imported as independent external work rather than used to prove itself. The only self-citation, [22], appears in a background list on the torsion function and is not load-bearing. The C2-regularity limitation is stated openly and does not smuggle in the conclusion. The skeptical concern about Lemma 1 is a logical gap in the proof of the quantifier over all vector fields, but the circularity score remains 0 because no step reduces by construction or by self-citation to the result being proved.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard tools in elliptic PDE theory: maximum principles, the isoperimetric inequality, the boundary point lemma, and the coarea formula. The variable b is a fixed input (the drift bound), not a fitted parameter. No new entities are introduced. The main domain assumption is C^2 regularity of Ω and C^2 regularity of the solution up to the boundary, which the paper cites from the fully nonlinear theory.

assumptions (6)
  • standard math Maximum principle for elliptic operators with drift: if -Δv + a·∇v ≥ 0 in Ω and v = 0 on ∂Ω, then v ≥ 0.
    Used in Lemma 1 to show u ≥ w.
  • standard math Isoperimetric inequality |∂Ω| ≥ c_d |Ω|^{1-1/d} with sharp constant c_d.
    Used in Lemmas 3 and 5 to replace |∂Ω| by c_d c^{1-1/d}.
  • standard math Zaremba-Hopf-Oleinik boundary point lemma: positive superharmonic functions in bounded C^2 domains have strictly positive inward normal derivative on the boundary.
    Used in Lemma 2 to ensure ∂u/∂n > 0, needed for the level-set volume expansion and Cauchy-Schwarz inequality.
  • standard math Coarea formula ∫_{u≤ε}|∇u| dx = ∫_0^ε H^{d-1}{u=t} dt.
    Used to compute the volume of Ω_ε and the boundary-layer gradient integral.
  • domain assumption Existence, uniqueness, and C^2 regularity of solutions of -Δu - b|∇u| = 1 in C^2 domains.
    The paper invokes Caffarelli-Cabre [13] for u ∈ C^2; this regularity underpins the level-set analysis.
  • domain assumption The domain Ω is C^2.
    Used throughout; the paper acknowledges the optimal regularity threshold is open (see intro and Lemma 2 discussion).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Trapping of Brownian Motion: A Nonlinear Analogue of the Torsion Function." pith.science (2026). https://pith.science/paper/VNZA6TEC

@misc{pith2026190806273,
  author       = {Pith},
  title        = {Pith review of: Optimal Trapping of Brownian Motion: A Nonlinear Analogue of the Torsion Function},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VNZA6TEC}},
  note         = {Machine review of arXiv:1908.06273}
}
abstract

We study the problem of maximizing the expected lifetime of drift diffusion in a bounded domain. More formally, we consider the PDE \[ - \Delta u + b(x) \cdot \nabla u = 1 \qquad \mbox{in}~\Omega\] subject to Dirichlet boundary conditions for $\|b\|_{L^{\infty}}$ fixed. We show that, in any given $C^2-$domain $\Omega$, the vector field maximizing the expected lifetime is (nonlinearly) coupled to the solution and satisfies $b = -\|b\|_{L^{\infty}} \nabla u/ |\nabla u|$ which reduces the problem to the study of the nonlinear PDE \[ -\Delta u - b \cdot \left| \nabla u \right| = 1,\] where $b = \|b\|_{L^{\infty}}$ is a constant. We believe that this PDE is a natural and interesting nonlinear analogue of the torsion function. We prove that, for fixed volume, $\| \nabla u\|_{L^1}$ and $\|\Delta u\|_{L^1}$ are maximized if $\Omega$ is the ball (the ball is also known to maximize $\|u\|_{L^p}$ for $p \geq 1$ from a result of Hamel \& Russ).

Figures

Figures reproduced from arXiv: 1908.06273 by the authors.

Figure 1
Figure 1. Brownian motion stays trapped for the longest time if it moves inside a ball and the vector field pushes it radially inside. 2010 Mathematics Subject Classification. 35B51, 49K20 (primary) and 60J60 (secondary). Key words and phrases. Drift Diffusion, Exit Time, Isoperimetric Inequality, Torsion function. The research of J.L. was supported in part by the National Science Foundation under award DMS-1454939. S.S. was … view at source ↗
Figure 2
Figure 2. Lemma 1 illustrated: the optimal vector field is orthog￾onal to the level sets. 3.2. An estimate on the L 1 norm of the gradient. The purpose of this section is to establish part of the statement of the main Theorem: among all domains Ω with fixed volume, we are interested in solutions of −∆u − kbkL∞|∇u| = 1 with Dirichlet boundary conditions. Among those solutions Z Ω |∇u|dx is maximized by the ball. This is the ma… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

  1. [13]

    Caffarelli and X

    L. Caffarelli and X. Cabre, Fully nonlinear elliptic equ ations, American Mathematical Society Colloquium Publications, 43, American Mathematical Socie ty, Providence, RI, 1995

  2. [1]

    Apushkinskaya and Alexander I

    D. Apushkinskaya and Alexander I. Nazarov, A counterexa mple to the Hopf–Oleinik lemma (elliptic case), Anal. PDE 9 (2016), p. 439 – 458

  3. [2]

    Baernstein II, Symmetrization in Analysis, Cambridg e University Press, 2019

    A. Baernstein II, Symmetrization in Analysis, Cambridg e University Press, 2019

  4. [3]

    Bandle, Isoperimetric inequalities and application s

    C. Bandle, Isoperimetric inequalities and application s. Monographs and Studies in Mathemat- ics, 7. Pitman (Advanced Publishing Program), Boston, Mass .-London, 1980

  5. [4]

    Ba˜ nuelos and T

    R. Ba˜ nuelos and T. Carroll, Brownian motion and the fund amental frequency of a drum, Duke Math. J. Volume 75, Number 3 (1994), 575–602

  6. [5]

    Ba˜ nuelos and T

    R. Ba˜ nuelos and T. Carroll, The maximal expected lifeti me of Brownian motion, Mathematical Proceedings of the Royal Irish Academy, 111 (2011), p.1–11

  7. [6]

    Ba˜ nuelos, M

    R. Ba˜ nuelos, M. van den Berg and T. Carroll, Torsional ri gidity and expected lifetime of Brownian motion. Journal of the London Mathematical Societ y 66 (2002), p. 499–512

  8. [7]

    Beck, The torsion function of convex domains of high ec centricity, Potential Analysis, to appear

    T. Beck, The torsion function of convex domains of high ec centricity, Potential Analysis, to appear

Show all 32 references
  1. [8]

    Beck, Uniform level set estimates for ground state eig enfunctions, SIAM J

    T. Beck, Uniform level set estimates for ground state eig enfunctions, SIAM J. Math. Anal. 50 (2018), p. 4483–4502

  2. [9]

    Burchard and M

    A. Burchard and M. Schmuckenschl¨ ager, Comparison theo rems for exit times, Geometric & Functional Analysis GAF A, 2001, p. 651–692

  3. [10]

    H. J. Brascamp and E. Lieb, On extensions of the Brunn-Mi nkowski and Prekopa-Leindler theorems, including inequalities for log concave function s, and with an application to a diffusion equation, J. Funct. Anal. 22 (1976), 366–389

  4. [11]

    Brasco and R

    L. Brasco and R. Magnanini, The Heart of a Convex Body, in :Geometric Properties for Parabolic and Elliptic PDE’s, Springer, 2013

  5. [12]

    Burdzy and W

    K. Burdzy and W. W erner, A counterexample to the ”hot spo ts” conjecture, Ann. Math. 149 (1999), p. 309–317

  6. [14]

    M. I. Freidlin and A. D. W entzell, Random pertubations o f dynamical systems. 3rd edition. Springer, New York, 2012

  7. [15]

    Freitas, Closed Nodal Lines and Interior Hot Spots of the Second Eigenfunctions of the Laplacian on Surfaces, Indiana Univ

    P. Freitas, Closed Nodal Lines and Interior Hot Spots of the Second Eigenfunctions of the Laplacian on Surfaces, Indiana Univ. Math. J. 51 (2002), 305 –316

  8. [16]

    Grieser and D

    D. Grieser and D. Jerison, Asymptotics of the first nodal line of a convex domain, Invent. Math., 125 (1996), p. 197–219. 11

  9. [17]

    Hamel, N

    F. Hamel, N. Nadirashvili and Y. Sire, Convexity of leve l sets for elliptic problems in convex domains or convex rings: Two counterexamples, Amer. J. Math 138 (2013), p. 499–527

  10. [18]

    Hamel and E

    F. Hamel and E. Russ. Comparison results and improved qu antified inequalities for semilinear elliptic equations, Math. Ann. 367 (2017), 311–372

  11. [19]

    Kawohl, Rearrangements and Convexity of Level Sets i n PDE, Lecture Notes in Mathe- matics 1150, Springer, Berlin, 1985

    B. Kawohl, Rearrangements and Convexity of Level Sets i n PDE, Lecture Notes in Mathe- matics 1150, Springer, Berlin, 1985

  12. [20]

    Kuran, On positive superharmonic functions in α -admissible domains, J

    U. Kuran, On positive superharmonic functions in α -admissible domains, J. London Math. Soc., (2) 29 (1984), no. 2, 269–275

  13. [21]

    Lions, Two geometrical properties of solutions o f semilinear problems, Appl

    P.-L. Lions, Two geometrical properties of solutions o f semilinear problems, Appl. Anal. 12 (1981), 267–272

  14. [22]

    Lu and S

    J. Lu and S. Steinerberger, A dimension-free Hermite-H adamard inequality via gradient es- timates for the torsion function (2019), arXiv preprint arX iv:1905.03216

  15. [23]

    Magnanini, An introduction to the study of critical p oints of solutions of elliptic and parabolic equations, Rend

    R. Magnanini, An introduction to the study of critical p oints of solutions of elliptic and parabolic equations, Rend. Istit. Mat. Univ. Trieste 48 (20 16), 121–166

  16. [24]

    L. G. Makar-Limanov, The solution of the Dirichlet prob lem for the equation ∆ u = −1 in a convex region, Mat. Zametki 9 (1971), 89–92 (Russian). Engl ish translation in Math. Notes 9 (1971), 52–53

  17. [25]

    Nazarov, A centennial of the Zaremba-Hopf-Oleinik l emma

    A. Nazarov, A centennial of the Zaremba-Hopf-Oleinik l emma. SIAM J. Math. Anal. 44 (2012), no. 1, 437–453

  18. [26]

    P´ olya, Torsional rigidity, principal frequeny, el ectrostatic capacity and symmetrization, Quart

    G. P´ olya, Torsional rigidity, principal frequeny, el ectrostatic capacity and symmetrization, Quart. Appl. Math. 6 (1948), p. 267–277

  19. [27]

    P´ olya and G

    G. P´ olya and G. Szeg˝ o, Isoperimetric Inequalities in Mathematical Physics, Annals of Math- ematics Studies, Princeton University Press, 1951

  20. [28]

    Pucci, J

    P. Pucci, J. Serrin, The maximum principle, Progress in Nonlinear Differential Equations and their Applications, 73. Birkhauser Verlag, Basel, 2007

  21. [29]

    Rachh and S

    M. Rachh and S. Steinerberger, On the location of maxima of solutions of Schroedinger’s equation, Comm. Pure. Appl. Math. 71 (2018), 1109–1122

  22. [30]

    Steinerberger, Topological bounds for Fourier coeffi cients and applications to torsion, Jour- nal of Functional Analysis 274 (2018), 1611–1630

    S. Steinerberger, Topological bounds for Fourier coeffi cients and applications to torsion, Jour- nal of Functional Analysis 274 (2018), 1611–1630

  23. [31]

    Talenti, Elliptic equations and rearrangements

    G. Talenti, Elliptic equations and rearrangements. An n. Scuola Norm. Sup. Pisa Cl. Sci. (4) 3 (1976), 697–718

  24. [32]

    Volkov, On differential properties of solutions to boundary value problems for the Laplace equation on polygons, Trudy Mat

    E.A. Volkov, On differential properties of solutions to boundary value problems for the Laplace equation on polygons, Trudy Mat. Inst. Steklov [Pro c. Steklov Inst. Math.], 77 (1965), 113–142. (Jianfeng Lu) Department of Mathematics, Department of Physics, and Depa rtment of Ch...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.