Evolution equations on time-dependent intervals

Athanasios S. Fokas; Baoqiang Xia; Beatrice Pelloni

REVIEW 3 major objections 5 minor 16 references

Evolution equations on time-dependent intervals

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

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proves that the missing boundary data for evolution equations on moving intervals can be recovered from a two-by-two system of integral equations.

desk verdict A competent extension of the unified transform to moving finite intervals, with a solid heat equation result and a Schrödinger theorem whose proof skips the hard convergence step. read the letter →

arxiv 1908.03729 v1 pith:BF4FAH3J submitted 2019-08-10 math.AP

Athanasios S. Fokas , Beatrice Pelloni , Baoqiang Xia This is my paper · ORCID

classification math.AP MSC 35K0535Q4135R3745D05

keywords linearevolutionequationstime-dependentintervalDirichlet-to-NeumannmapheatequationSchrödingerVolterraintegralglobalrelationmovingboundary

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 claims that for linear evolution equations on a finite interval whose endpoints move in time, the missing boundary data—the spatial derivatives of the solution at the two moving endpoints—are determined uniquely by the prescribed initial and Dirichlet data. For the heat equation the claim holds for arbitrary differentiable, non-crossing boundary curves; the unknown Neumann values solve a system of two linear Volterra integral equations with explicit kernels. For the linear Schrödinger equation the same characterization holds only when the boundary curves satisfy a convexity condition or are linear with compatible slopes. If true, this completes the Dirichlet-to-Neumann construction for these problems and yields an explicit solution representation on the moving domain. Moving-boundary problems arise in melting, diffusion, and interfacial dynamics, where the evolution of boundary fluxes is the central unknown.

What carries the argument

The key machinery is the global relation (2.8), a complex-plane integral identity from the unified transform method: with $Q$ defined in (2.3), the PDE is written in divergence form, Green’s theorem yields a relation among the Fourier transforms of the initial data, the solution on the interval, and boundary integrals. For each equation, multiplying by $i\lambda$ and integrating by parts converts the global relation into an equation for the Fourier transform of $q_x$, which is then inverted and evaluated at the two moving endpoints. The result is a system of two linear Volterra equations; the kernels $K_{jm}$ in (1.9) and (1.12)–(1.13) carry the interaction between the endpoints, with the diagonal case $j=m$ weakly singular and the off-diagonal cases regular for heat or regularized by $\varepsilon$ for Schrödinger. The $\varepsilon$-regularized kernels and the geometric denominators $H_1,H_2$ are what make the existence proof work.

What would settle it

Take a boundary pair satisfying (1.14), for example $l_1(t)=-t+t^2/2$ and $l_2(t)=1+t-t^2/2$ on a small interval, solve the regularized system (4.1) numerically for decreasing $\varepsilon>0$, and check whether the solutions converge to a limit satisfying (4.4)–(4.5). If the limit fails to exist or violates the equations, the Schrödinger theorem’s proof has a hole; for the heat equation, testing (1.7) against an exact solution with known boundary fluxes would confirm or falsify the uniqueness claim.

Watch

Extended reading notes

Core claim

The central discovery is that the global relation—an identity obtained by applying Green’s theorem to the divergence form of the PDE—contains enough information to close the Dirichlet-to-Neumann map on a finite time-dependent interval. Evaluating the inverse Fourier transform of the derivative $q_x(x,t)$ at the two endpoints yields a coupled system, equation (1.7) for the heat equation and (1.10) for the linear Schrödinger equation. For heat, the coupling kernels are weakly singular at worst, and classical Volterra theory gives a unique $C^1$ solution. For Schrödinger, the cross kernels are genuinely singular and must be regularized; after an $\varepsilon$-limit and integration by parts, the paper obtains a generalized Volterra system, equations (4.4)–(4.5), whose kernels are weakly singular provided the geometric quantities $H_1$ and $H_2$ do not vanish, which is guaranteed by the convexity condition (1.14) or by linear boundaries of the form (4.6).

Load-bearing premise

For the Schrödinger theorem, the proof assumes that the solutions of the regularized approximating system converge as the regularization is removed, without actually proving that convergence for the two-equation system.

Editorial extensions

If this is right

For the heat equation on a time-dependent interval, the Dirichlet-to-Neumann map is well defined and unique: the boundary fluxes $q_x(l_1(t),t)$ and $q_x(l_2(t),t)$ are the unique $C^1$ solution of the system (1.7), with no extra geometric condition on the boundaries beyond differentiability and $l_1<l_2$.
For the linear Schrödinger equation, the same conclusion holds for boundary curves satisfying the convexity condition (1.14) or for linear boundaries of the form (4.6), but the paper does not assert it for arbitrary moving boundaries.
Once the Neumann data are obtained, formula (2.9) gives an explicit representation of the solution on the whole domain, and the known uniqueness result for the global relation guarantees that this representation is the actual unique solution of the boundary value problem.
The result extends the earlier half-line analysis to the two-sided moving interval, where the main new feature is a coupled system of two integral equations rather than a single one, and the off-diagonal kernels require regularization in the Schrödinger case.

Reading between the lines

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

An implicit extension is that the same global-relation-plus-Volterra strategy should apply to other linear evolution equations on time-dependent intervals, with the main difficulty shifting to the oscillatory regularization of the cross kernels.
The convexity condition (1.14) is sufficient but likely not necessary; the actual requirement is only that $H_1(t,s)$ and $H_2(t,s)$ stay nonzero, so the theorem could be pushed to broader boundary classes by proving non-vanishing under weaker hypotheses.
A practical consequence the paper does not spell out is numerical: the explicit system (1.7) or (4.4)–(4.5) can be discretized directly, giving a way to compute boundary fluxes on moving domains without resolving the PDE in the interior.

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.

Desk Editor's Note

read the letter

This paper extends the Fokas unified transform to linear evolution PDEs on a time-dependent interval with two moving boundaries. The heat equation part is essentially complete: the unknown Neumann data are characterized as the unique solution of a system of two Volterra integral equations with explicit, weakly singular kernels, and a classical existence theorem applies. That is a genuine step beyond the half-line analysis of Fokas–Pelloni (2012). The Schrödinger case is where I have real reservations.

The formal derivation of the integral equations from the global relation is coherent, and Claims 1–4 contain the honest Gaussian computations. The regularization of the strongly singular kernels via ε is a sensible device, and the limiting equations (4.4)–(4.5) are formally right under the geometric condition (1.14) or the linear-boundary condition (4.6). The citation pattern is reasonable: uniqueness is delegated to Xia (2019), and the Volterra machinery to Miller–Feldstein, which is acceptable for a sketch.

The soft spot is Theorem 2. The proof shows convergence of the kernels as ε → 0, but the existence of the limits of the solutions (f^ε, g^ε) is never established. The sentence 'Hence for f1(t)=lim f^ε(t), g1(t)=lim g^ε(t) ...' is exactly where the argument jumps. The stress-test note is right: the vector extension is not immediate. The limiting equations contain f′ and g′, so even uniform convergence of f^ε would not pass to the derivative terms without a Hölder or C^1 bound. The ε-regularized system involves kernels with L^1 norms that may grow like ε^{-1/4} near the diagonal, so a naive resolvent bound diverges. You need a compactness/equicontinuity argument, or an ε-uniform Gronwall estimate, to justify the limits. This is not a cosmetic gap; it is a missing step in the proof of the main Schrödinger result.

I should also note that Theorem 1's proof is a reference to earlier work, not self-contained, but that is standard and acceptable. The heat equation part is solid.

Who is this for? People working on the Fokas method and moving-boundary problems. They will find the heat equation result useful and the Schrödinger formulation a promising starting point, but they should treat Theorem 2 as conditional pending a complete proof.

I would send it to peer review. A serious referee could help the authors fill the gap or see that the claim is actually false. The paper deserves engagement, not a desk reject. I would cite the heat equation result, but not the Schrödinger theorem as proven until the ε→0 limit is sorted out.

Referee Report

3 major / 5 minor

Summary. The paper develops a Fokas-type (unified transform) analysis for linear evolution PDEs on a time-dependent interval l1(t)<x<l2(t), 0<t<T. The authors derive a global relation from a divergence form and invert the Fourier transform to obtain formal integral representations of the unknown Neumann boundary values. For the heat equation they obtain a system of two linear Volterra integral equations, (1.7), with explicitly computed Gaussian kernels, and prove in Theorem 1 that this system has a unique C^1 solution. For the linear Schrödinger equation they obtain a system (1.10) containing an ε-regularized kernel; Theorem 2 asserts that, under the convexity condition (1.14) or the linear-boundary condition (4.6), the ε-limit exists and yields a generalized Volterra system (4.4)-(4.5) with a unique solution. The paper also sketches the reduction of the general evolution PDE case to the global relation and cites prior work for uniqueness of the resulting initial-boundary value problem.

Significance. If the main results are correct, the paper gives an explicit characterization of the Dirichlet-to-Numann map for the heat and linear Schrödinger equations on time-dependent intervals, extending earlier half-line results of Fokas-Pelloni and Xia. The heat-equation part appears well supported: the global-relation derivation is self-contained and the kernel computations in Claims 1-2 are standard, with the weakly singular Volterra theory invoked appropriately. The claimed Schrödinger result is more delicate because the limiting kernels are obtained through an ε-regularization and the limiting equations contain first derivatives of the unknown functions. The paper also honestly delineates the restrictive class of boundary curves for which the LS result is claimed. Because the proof of Theorem 2 has a genuine gap in the ε-limit passage, the significance of the Schrödinger characterization rests on an unproven step; the heat-equation contribution stands independently.

major comments (3)

[§4.2, Eq. (4.1)-(4.5)] The proof of Theorem 2 does not establish the existence of the limits f1(t)=lim f^ε_1(t) and g1(t)=lim g^ε_1(t) that are used to pass from the regularized system (4.1) to the limiting system (4.4)-(4.5). The text proves convergence of the kernels for a fixed smooth h by dominated convergence, but no argument is given for convergence of the solutions of the coupled Volterra system; the statement that the vector extension is 'immediate' is not supported. This is load-bearing, since the characterization of the unknown Neumann values for the Schrödinger equation is exactly the unique solvability of (4.4)-(4.5).
[§4.2, scalar reduction and derivative terms] Even if one had uniform convergence of f^ε and g^ε, the limiting equations (4.4)-(4.5) contain f'_1 and g'_1 in the integrands, so uniform convergence alone would not justify passing the ε-limit through the derivative terms. A Hölder or C^1 bound, or an alternative compactness argument, is required. The manuscript does not supply such a bound, and the scalar-case reduction cannot address the coupling between the two components.
[§4.2, existence/uniqueness of the generalized Volterra system] The claim that (4.4)-(4.5) is a 'generalised Volterra integral equation of the second kind with a weakly integral kernel' and that a modification of the iterative proof works is only supported by a reference to Brunner. The hypotheses of the relevant existence theorem are not verified: the equations are not in standard Volterra form because of the derivative terms, and the kernels contain H1(t,s)^{-1} and H2(t,s)^{-1}, whose regularity is only shown under (1.14) or (4.6). The proof should either state and verify a precise theorem from Brunner or provide a self-contained fixed-point argument.

minor comments (5)

[Theorem 2 statement] There is a typo: 'Assume the the boundary functions' should read 'Assume the boundary functions'.
[§4.2, H1 estimate] In the inequality proving H1(t,s)<0, the term 'l′(s)' should be 'l′_2(s)' for consistency with the definition (4.3).
[Equation (1.12)] In the definition of K_jj, the symbol 'ε > 0' appears, but K_jj does not depend on ε; the condition should probably be '0<s<t<T' only.
[Remark 2] The phrase 'Theorem (2)' should be 'Theorem 2'.
[Section 4.2, text near (4.2)] The notation E12(t,s,ε) is introduced for the exponential, but the subsequent display uses E12(t,s) without ε in some places; the dependence on ε should be tracked consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the Neumann-data equations are derived from the global relation with explicit kernels, and prior-work citations provide external support rather than restating the target theorem.

full rationale

The derivation of the Dirichlet-to-Neumann characterization is self-contained: equations (1.7) and (1.10) are obtained by applying Fourier inversion to the global relation (2.8)/(3.2)/(3.19) and evaluating at the boundaries; the kernels K_jm and forcing terms N_j are computed explicitly from the prescribed data q0, f0, g0 and the boundary curves. No parameter is fitted, and no conclusion is inserted by definition: f1 and g1 genuinely appear as unknowns in coupled Volterra equations solved by classical theory. The paper appeals to [Fokas-Pelloni(2012)] for the analogous half-line analysis and to [Xia(2019)] for uniqueness of the BVP solution; these are published external results covering different (single-boundary or global-relation) facts, not restatements of Theorems 1-2, so they are not load-bearing circularity. The only flagged weakness is in Theorem 2: the epsilon-to-zero limit is justified for scalar kernels, while convergence of the coupled solution (f^epsilon, g^epsilon) is asserted with 'the extension ... is immediate', and no uniform boundedness or compactness is supplied. That is a completeness or correctness gap, not a circularity, because the limiting equations (4.4)-(4.5) are not assumed as inputs. Hence no circular step can be exhibited, and the correct score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The derivation introduces no free parameters; the boundary curves and data are given inputs. The ε-regularized kernels are a technical device, not a new physical entity. The main load-bearing inputs are standard Fourier/Volterra theory, the zero-extension boundary convention, the external uniqueness result of Xia (2019), and for the Schrödinger theorem the geometric condition (1.14).

assumptions (5)

standard math The divergence form (2.4) and Green's theorem (2.5) are valid for the domain Ω(T).
Used to derive the global relation (2.8), the starting point for all integral representations.
domain assumption q and qx can be extended by zero outside [l1(t),l2(t)] so that Fourier inversion applies at the boundary points with a factor 1/π rather than 1/(2π).
This is the standard 'jump at the endpoint' convention in the Fokas method; it is asserted in Section 3 before equations (3.8)-(3.9).
standard math Classical Volterra existence/uniqueness theory for weakly singular kernels applies to the systems (1.7) and (4.4)-(4.5).
Invoked in Section 4.1 via [Miller-Feldstein(1971)] and in Section 4.2 via [Brunner(2017)].
domain assumption The external result [Xia(2019)] that sufficiently regular data satisfying the global relation yield the unique solution of the boundary value problem.
Used in Section 2 (pages 7-8) to assert that the characterized Neumann values indeed correspond to the true solution.
domain assumption For Theorem 2, the boundary curves satisfy l1'<0, l1''≥0, l2'>0, l2''≤0 (condition (1.14)).
This convexity condition is used in Section 4.2 to prove H1 and H2 never vanish, which is needed for the dominated convergence argument.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Evolution equations on time-dependent intervals." pith.science (2026). https://pith.science/paper/BF4FAH3J

@misc{pith2026190803729,
  author       = {Pith},
  title        = {Pith review of: Evolution equations on time-dependent intervals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BF4FAH3J}},
  note         = {Machine review of arXiv:1908.03729}
}

abstract

We study initial boundary value problems for linear evolution partial differential equations (PDEs) posed on a time-dependent interval $l_1(t)<x<l_2(t)$, $0<t<T$, where $l_1(t)$ and $l_2(t)$ are given, real, differentiable functions, and $T$ is an arbitrary constant. For such problems, we show how to characterise the unknown boundary values in terms of the given initial and boundary conditions. As illustrative examples we consider the heat equation and the linear Schr\"{o}dinger equation. In the first case, the unknown Neumann boundary values are expressed in terms of the Dirichlet boundary values and of the initial value through the unique solution of a system of two linear integral equations with explicit kernels. In the second case, a similar result can be proved but only for a more restrictive class of boundary curves.}

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages

[1]

Brunner, Volterra integral equations: an introduction to theory and applications , Cambridge University Press, 2017

H. Brunner, Volterra integral equations: an introduction to theory and applications , Cambridge University Press, 2017

work page 2017
[2]

Fokas, Proc

A.S. Fokas, Proc. Roy. Soc. London Ser. A, 453 (1997) 1411-1443

work page 1997
[3]

Fokas, J

A.S. Fokas, J. Math. Phys., 41 (2000) 4188-4237

work page 2000
[4]

Fokas and B

A.S. Fokas and B. Pelloni, Stud. Appl. Math., 129 (2012) 51-90

work page 2012
[5]

De Lillo and A.S

S. De Lillo and A.S. Fokas, Inv. Prob., 23 (2007) 1699

work page 2007
[6]

Fokas and B

A.S. Fokas and B. Pelloni, J. Math. Phys. 48 (2007) 013502

work page 2007
[7]

Pelloni, Inv

B. Pelloni, Inv. Prob., 24 (2008) 015004

work page 2008
[8]

Feldstein, SIAM J

R.K.Miller and A. Feldstein, SIAM J. Math. An. 2 (1971) 242-258

work page 1971

Show all 16 references

[9]

Xia, The Fokas method for integrable evolution equations on a time-dependent interval, J

B. Xia, The Fokas method for integrable evolution equations on a time-dependent interval, J. Math.Phys. 60 (2019) 023504

2019
[10]

Fokas, A Unified Approach to Boundary Value Problems , Society for Industrial and Applied Mathematics (SIAM), Philadelphia, 2008, 27(2)

A.S. Fokas, A Unified Approach to Boundary Value Problems , Society for Industrial and Applied Mathematics (SIAM), Philadelphia, 2008, 27(2)

2008
[11]

Deconinck, T

B. Deconinck, T. Trogdon, and V. Vasan, SIAM Rev., 56 (2014) 159-186

2014
[12]

Deconinck, Q

B. Deconinck, Q. Guo, E. Shlizerman, V. Vasan, Quart. Appl. Math. (2017)

2017
[13]

Fokas and B

A.S. Fokas and B. Pelloni, Phys. Rev. Lett., 84 (2000) 4785-4789

2000
[14]

Fokas and B

A.S. Fokas and B. Pelloni, Inv. Prob., 17 (2001) 919-935

2001
[15]

Fokas and P.F

A.S. Fokas and P.F. Schultz, Comm. Pure Appl. Math. 56 (2003) 517-548

2003
[16]

Fokas, Introduction and applications of complex variables, Cambridge University Press, 2nd ed

M.J Ablowitz and A.S. Fokas, Introduction and applications of complex variables, Cambridge University Press, 2nd ed. 2004

2004

Pith tools

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

[1] [1]

Brunner, Volterra integral equations: an introduction to theory and applications , Cambridge University Press, 2017

H. Brunner, Volterra integral equations: an introduction to theory and applications , Cambridge University Press, 2017

work page 2017

[2] [2]

Fokas, Proc

A.S. Fokas, Proc. Roy. Soc. London Ser. A, 453 (1997) 1411-1443

work page 1997

[3] [3]

Fokas, J

A.S. Fokas, J. Math. Phys., 41 (2000) 4188-4237

work page 2000

[4] [4]

Fokas and B

A.S. Fokas and B. Pelloni, Stud. Appl. Math., 129 (2012) 51-90

work page 2012

[5] [5]

De Lillo and A.S

S. De Lillo and A.S. Fokas, Inv. Prob., 23 (2007) 1699

work page 2007

[6] [6]

Fokas and B

A.S. Fokas and B. Pelloni, J. Math. Phys. 48 (2007) 013502

work page 2007

[7] [7]

Pelloni, Inv

B. Pelloni, Inv. Prob., 24 (2008) 015004

work page 2008

[8] [8]

Feldstein, SIAM J

R.K.Miller and A. Feldstein, SIAM J. Math. An. 2 (1971) 242-258

work page 1971