Determining the optimal coefficient of the spatially periodic Fisher-KPP equation that minimizes the spreading speed

Ryo Ito

REVIEW 4 major objections 3 minor 17 references

Determining the optimal coefficient of the spatially periodic Fisher-KPP equation that minimizes the spreading speed

T0 review · 4 major / 3 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 For a periodic Fisher–KPP habitat, the slowest possible front speed equals the lower bound exactly when $r/\langle r\rangle_a+\langle d\rangle_h/d=2$, and the speed-minimizing growth rate is $r_d(x)=\alpha(2-\langle d\rangle_h/d(x))$.

desk verdict The main theorem is false: the constant test function is not a minimizer under the paper's own condition, and Example 1 gives a speed strictly larger than the claimed value. read the letter →

arxiv 1908.09538 v1 pith:6B5SV5HS submitted 2019-08-26 math.AP

Ryo Ito This is my paper · ORCID

classification math.AP MSC 35K9135C0792D40

keywords Fisher-KPPequationminimalwavespeedspreadingspatiallyperiodiccoefficientsprincipaleigenvaluevariationalcharacterizationoptimalgrowthrateharmonicmean

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

The paper establishes when the minimal speed of a one-dimensional spatially periodic Fisher–KPP equation coincides with the simple homogenized lower bound $2\sqrt{\langle d\rangle_h\langle r\rangle_a}$, where $\langle r\rangle_a$ is the spatial mean of the growth rate and $\langle d\rangle_h$ is the harmonic mean of the diffusion. It proves this happens exactly when the pointwise ratio condition $r/\langle r\rangle_a+\langle d\rangle_h/d=2$ holds, and in that case the speed is explicitly calculable. It then solves the natural inverse problem: among all growth rates with a fixed positive mean $\alpha$, the speed is minimized by $r_d(x)=\alpha(2-\langle d\rangle_h/d(x))$, a function that is large where diffusion is large and small where diffusion is small. These results give the first exact, nonconstant example of the minimal (spreading) speed for spatially periodic coefficients, and they show that the shape of environmental heterogeneity, not just its average, controls the invasion speed.

What carries the argument

The object carrying the argument is the variational formula for the principal eigenvalue of the linearized operator: $k_\lambda(d,r)=\min_{\phi\in E_L} I(\phi;\lambda,d,r)$, where $E_L$ is the set of positive $L$-periodic profiles with $L^2$ norm one and $I(\phi;\lambda,d,r)=\int_0^L d|\phi'|^2-\int_0^L r\phi^2-\lambda^2 L^2\int_0^L (1/d)\phi^2$. Combined with the min-max formula $c^*_d(r)=\min_{\lambda>0}(-k_\lambda(d,r)/\lambda)$, this functional converts the speed into a variational question. The proof's hinge is the constant test profile $\phi_0\equiv 1/\sqrt{L}$: equality in the lower bound forces $\phi_0$ to be the actual minimizer at $\lambda_0=\sqrt{\langle r\rangle_a/\langle d\rangle_h}$, and inserting it into the Euler–Lagrange equation reduces exactly to the relation $r/\langle r\rangle_a+\langle d\rangle_h/d=2$. That relation is then solved explicitly for $r$, producing the optimizer.

What would settle it

Take the period-$2\pi$ coefficients $r(x)=1+\tfrac12\sin x$ and $d(x)=1/(1-\tfrac12\sin x)$, which satisfy the ratio condition; if a direct numerical evaluation of the principal-eigenvalue formula gave a minimal speed different from $2$, the equality characterization would be false.

Watch

Extended reading notes

Core claim

The central claim is that equality in the universal lower bound for the minimal speed is characterized by a pointwise ratio condition: $c^*_d(r)=2\sqrt{\langle d\rangle_h\langle r\rangle_a}$ holds if and only if $r(x)/\langle r\rangle_a+\langle d\rangle_h/d(x)=2$ for all $x$. Under this condition the unique critical parameter is $\lambda_0=\sqrt{\langle r\rangle_a/\langle d\rangle_h}$, the unique minimizer of the variational functional is the constant $\phi_0\equiv 1/\sqrt{L}$, and the minimal speed is explicitly $2\sqrt{\langle d\rangle_h\langle r\rangle_a}$. Solving the ratio condition for $r$ gives the optimizer of the constrained minimization problem: $r_d(x)=\alpha(2-\langle d\rangle_h/d(x))$ for a prescribed mean $\alpha$, and the same function minimizes the speed for every positive rescaling of $d$. The paper therefore provides, as it states, the first exact value of the minimal speed for a spatially periodic Fisher–KPP equation with nonconstant coefficients.

Load-bearing premise

The proof that equality forces the ratio condition assumes that the constant profile $\phi_0\equiv 1/\sqrt{L}$ is a genuine minimizer of the variational functional at $\lambda_0$, so that evaluating the Euler–Lagrange equation at $\phi_0$ is legitimate; if another profile gave a smaller value, the argument would not go through.

Editorial extensions

If this is right

In any habitat satisfying $r/\langle r\rangle_a+\langle d\rangle_h/d=2$, the exact invasion speed is $2\sqrt{\langle d\rangle_h\langle r\rangle_a}$; Example 1 computes $c^*=2$ for $r=1+\tfrac12\sin x$ and $d=1/(1-\tfrac12\sin x)$.
The minimizing growth rate for prescribed mean $\alpha$ is $r_d(x)=\alpha(2-\langle d\rangle_h/d(x))$, so the slowest spread occurs when growth concentrates where diffusion is strongest.
Because $r_{kd}=r_d$ for every $k>0$, rescaling the diffusion amplitude does not change the optimal growth shape; only the normalized shape of $d$ matters.
If the ratio condition fails, the strict inequality $c^*_d(r)>2\sqrt{\langle d\rangle_h\langle r\rangle_a}$ holds; when $r$ is fixed and $2-r/\langle r\rangle_a$ is not a positive diffusion, no admissible $d$ can attain the lower bound.
Under the equality condition, the principal eigenfunction of the linearized operator is constant if and only if $d$ is constant, so the nonconstant optimal regime genuinely has a nonconstant eigenfunction.

Reading between the lines

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

Editorial extension: The paper allows sign-changing $r$; if one imposes the biological constraint $r\ge 0$, the variational problem changes and the explicit optimizer may no longer be admissible, a case the paper does not treat.
Editorial extension: Because $r_{kd}=r_d$ for every rescaling $k>0$, the optimization is really over the normalized shape of $d$; one could equally fix $r$ and ask which diffusion shape minimizes the speed, a reverse question the present theorem does not answer.
Editorial extension: The equality lower bound is also the homogenization limit $L\to 0$, so the ratio condition can be read as a zero-correction regime; it would be natural to test whether similar exact-speed relations hold for time-periodic or multi-dimensional analogues.

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

The central theorem is wrong. Under the paper's own condition (1.3), the constant function φ0 = 1/√L is not a minimizer of Nadin's variational functional; the first variation in a mean-zero direction is nonzero for any nonconstant r. Example 1 is a concrete counterexample: the claimed speed c* = 2 is not attained, and the actual speed is larger. In addition, the Cauchy-Schwarz step in the proof of (2)⇒(1) asserts an inequality that is generally false. The paper also has an internal inconsistency: the printed Nadin functional includes a factor L² in the last term, which contradicts the author's own evaluation in the proof of (1.1) and hides the variational error.

That said, the question is a good one and the explicit candidate minimizer r_d(x) = α(2 − ⟨d⟩h/d(x)) is a natural guess. The author correctly identifies the origin of condition (1.3) in El Smaily-Hamel-Roques [5] and cites the relevant literature honestly. The proof of the lower bound (1.1) is standard and correct.

The soft spots are not minor; they are load-bearing. The equivalence in Theorem 2.6 collapses, and with it the claimed explicit value of the speed. A reader working on periodic KPP speed minimization might keep the candidate formula as a conjecture, but the present proof does not support it. The paper should not be published. I would not send it to referees; the counterexample is simple enough to verify directly.

Referee Report

4 major / 3 minor

Summary. The paper studies the periodic Fisher-KPP equation u_t = (d(x)u_x)_x + (r(x)-u)u and considers the problem of minimizing the minimal wave speed c*_d(r) over growth rates r with fixed spatial mean, for a given periodic diffusion coefficient d. The main result, Theorem 2.6, asserts that c*_d(r) equals the lower bound 2 sqrt(<d>_h <r>_a) if and only if the coefficient pair satisfies r/<r>_a + <d>_h/d = 2. From this, the paper derives an explicit minimizer r_d(x) = alpha (2 - <d>_h/d(x)) for the minimization problem (P)_d and presents examples, claiming the first exact calculable minimal speed for a spatially periodic Fisher-KPP equation. The proof uses Nadin's variational characterization of the principal eigenvalue k_lambda(d,r) and an Euler-Lagrange analysis of a functional built from that formula.

Significance. The question addressed is natural and the intended result -- an explicit optimizer for the spreading speed under a mean constraint -- would be a valuable contribution if correct. The paper correctly recalls Nadin's variational formula and the known homogenization condition from [5], and the lower-bound proof in Section 3.2 is a standard and correct application of a constant test function. However, the central equivalence in Theorem 2.6 is not established; in fact, the paper's own Example 1 provides a concrete coefficient pair satisfying (1.3) for which the claimed equality c*_d(r)=2 fails. The flaw lies in the assertion that the constant function phi_0 minimizes the variational functional, which is false for nonconstant r. Consequently, the explicit minimizer and the claimed 'first calculable example' are unsupported, and the main contribution as stated does not hold.

major comments (4)

[Section 2.2 / Proposition 2.3] The functional I(phi;lambda,d,r) is defined with the term -lambda^2 L^2 integral_0^L (1/d) phi^2 dx, but the evaluation at phi_0 = 1/sqrt(L) in the proof of (1.1) (Section 3.2) and in equation (3.4) uses I(phi_0) = -<r>_a - lambda^2 <d>_h. These are incompatible: with the printed definition, integral_0^L (1/d) phi_0^2 dx = (1/L) integral_0^L (1/d) dx = 1/<d>_h, so I(phi_0) = -<r>_a - lambda^2 L^2 / <d>_h, not the stated value. The correct Nadin functional presumably has the term -lambda^2 / (integral (1/d) phi^2 dx); with that correction the evaluation is consistent, but then the subsequent Euler-Lagrange step in Section 3.3 does not match the printed formula. This inconsistency propagates through the proof of Theorem 2.6.
[Section 3.3, proof of Theorem 2.6, (1)=>(2)] The proof asserts after (3.1) that the constant function phi_0 = 1/sqrt(L) minimizes the functional I(.;lambda_0,d,r) under condition (1.3). This is false. For the correct Nadin functional, the first variation at phi_0 in any mean-zero direction psi is (2/sqrt(L)) integral (lambda_0^2 <d>_h^2/d - r) psi dx, which under (1.3) equals (4/sqrt(L)) integral (<r>_a - r) psi dx. This expression is nonzero whenever r is nonconstant, so phi_0 is not a stationary point and cannot be the minimizer. The Euler-Lagrange substitution leading to r/<r>_a + <d>_h/d = 2 is therefore invalid. In Example 1, taking psi = sin x gives a strictly negative first variation, so k_{lambda_0}(d,r) < -2<r>_a and c*_d(r) > 2 sqrt(<d>_h <r>_a), contradicting the claimed equality (1).
[Section 3.3, proof of Theorem 2.6, (2)=>(1)] The Cauchy-Schwarz step in the paragraph following (3.5) is incorrect. The inequality is printed as (integral (1/d) phi^2 dx)^2 >= (integral (1/d) dx)^2, which does not follow from Cauchy-Schwarz and fails, for instance, for phi = phi_0 because then the left-hand side equals (1/L^2)(integral (1/d) dx)^2, smaller than the right-hand side when L>1. The subsequent bound also implicitly requires the inequality <d>_h >= L, which is not an assumption of the problem. Consequently, the upper bound c*_d(r) <= 2 sqrt(<d>_h <r>_a) is not established.
[Section 3.3, Theorem 2.10] Two proofs are labeled 'Proof of Theorem 2.10'; the first of them actually proves Corollary 2.8 (the assertion about r_{kd}). Beyond this presentational duplication, the proof of Theorem 2.10 invokes the equality c*_d(r) = 2 sqrt(<d>_h <r>_a) as an assumption, but that equality is exactly the unproved (in fact false) statement of Theorem 2.6. Thus Theorem 2.10, while conditional in form, does not provide independent support for the main claims.

minor comments (3)

[Section 3.3, Euler-Lagrange equation] In the displayed Euler-Lagrange equation after the variation computation, the second integral is printed as -integral phi psi dx; it should be -integral r phi psi dx. The factor r is missing.
[Equation (3.2)] The displayed equation (3.2) is garbled: the fraction involving lambda_0^2 L^2 / 3 and the placement of d(x) in the denominator cannot be read unambiguously. Please rewrite it carefully because it is the step that allegedly produces (1.3).
[Throughout] The term 'first calculable example' in the abstract and Section 3.1 is overstated even apart from the proof error: the example's speed is calculated through the theorem rather than independently, and the condition (1.3) is already present in [5] as a homogenization criterion. The novelty should be framed more carefully.

Circularity Check

1 steps flagged · score 1.0 of 10

Minor tautology in Theorem 2.6(3), but the main equivalence (1)⇔(2) is not circular.

self definitional [Theorem 2.6, item (3), and §3.3 proof of (1)⇔(3)]
"Set λ0 = sqrt(⟨r⟩a/⟨d⟩h) and ϕ0≡1/√L. Then c∗d(r) = −λ−1 0 I(ϕ0;λ0,d,r). (2.6) ... I(ϕ0) =−⟨r⟩a−λ2 0⟨d⟩h =−2⟨r⟩a."

By the paper's own definitions, I(φ0;λ0,d,r) = -⟨r⟩a - λ0²⟨d⟩h = -2⟨r⟩a and λ0 = sqrt(⟨r⟩a/⟨d⟩h), so the right-hand side of (2.6) is 2⟨r⟩a/λ0 = 2√(⟨d⟩h⟨r⟩a), which is exactly the equality stated in item (1). Thus (3) is a restatement of (1) by construction; the Euler-Lagrange argument offered for (3)⇒(1) adds no independent content. This tautology is not load-bearing for the main equivalence (1)⇔(2), which is derived from external variational formulas attributed to Nadin and Berestycki-Hamel-Nadirashvili.

full rationale

The central derivation chain is not circular. The lower bound c*_d(r) ≥ 2√(⟨d⟩h⟨r⟩a) is quoted from Nadin via Proposition 2.3, an external variational characterization, and the Berestycki-Hamel-Nadirashvili formula is also an external reference. The paper fits no parameters and does not use its own conclusion as an input. The condition r/⟨r⟩a + ⟨d⟩h/d = 2 is imported from El Smaily-Hamel-Roques [5], not from the author's own work, and is then used in Theorem 2.6 to prove equality in the lower bound and to construct the minimizer rd(x) = α(2 - ⟨d⟩h/d(x)). No load-bearing self-citation occurs; the only self-citation [7] appears in a 'see also' list and is not used in the proofs. The one genuinely circular-seeming element is statement (3) of Theorem 2.6, which is just a notational restatement of statement (1): substituting the definitions gives the same equality. This is a presentational redundancy rather than a derivation from an assumed conclusion, and it does not affect the paper's main claim. Mathematical-correctness concerns about the Euler-Lagrange computation and the Cauchy-Schwarz step are separate from circularity and are not scored here.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claim rests entirely on known variational characterizations of the principal eigenvalue and the minimal speed. No new entities or fitted parameters are introduced. The flaw is not an unjustified axiom but an incorrect deduction from the cited variational formula.

assumptions (4)

domain assumption Nadin's variational formula for the principal eigenvalue k_λ(d,r) (Proposition 2.3)
The whole proof relies on this representation; it is an external theorem from [13].
domain assumption Berestycki-Hamel-Nadirashvili formula c* = min_{λ>0}(-k_λ/λ) (equation 2.4)
Established in [2]; used to convert eigenvalue estimates into speed estimates.
domain assumption Existence of the minimal wave speed and its equality with the spreading speed
From [17] and [2]; the paper takes this as background.
domain assumption The standing assumptions inf d > 0 and <r>_a > 0 (2.1)
Needed for the existence of positive periodic stationary states and spreading speeds.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Determining the optimal coefficient of the spatially periodic Fisher-KPP equation that minimizes the spreading speed." pith.science (2026). https://pith.science/paper/6B5SV5HS

@misc{pith2026190809538,
  author       = {Pith},
  title        = {Pith review of: Determining the optimal coefficient of the spatially periodic Fisher-KPP equation that minimizes the spreading speed},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6B5SV5HS}},
  note         = {Machine review of arXiv:1908.09538}
}

abstract

This paper is concerned with the spatially periodic Fisher-KPP equation $u_t=(d(x)u_x)_x+(r(x)-u)u$, $x\in \mathbb{R}$, where $d(x)$ and $r(x)$ are periodic functions with period $L>0$. We assume that $r(x)$ has positive mean and $d(x)>0$. It is known that there exists a positive number $c^*_d(r)$, called the minimal wave speed, such that a periodic traveling wave solution with average speed $c$ exists if and only if $c \geq c^*_d(r)$. In the one-dimensional case, the minimal speed $c^*_d(r)$ coincides with the ``spreading speed'', that is, the asymptotic speed of the propagating front of a solution with compactly supported initial data. In this paper, we study the minimizing problem for the minimal speed $c^*_d(r)$ by varying $r(x)$ under a certain constraint, while $d(x)$ arbitrarily. We have been able to obtain an explicit form of the minimizing function $r(x)$. Our result provides the first calculable example of the minimal speed for spatially periodic Fisher-KPP equations as far as the author knows.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 17 canonical work pages

[5]

ElSmaily, F

M. ElSmaily, F. Hamel and L. Roques, Homogenization and inﬂuence of frag- mentation in a biological invasion model , Disc. Cont. Dyn. Syst. A, 25: 321-342, 2009

work page 2009
[1]

Berestycki and F

H. Berestycki and F. Hamel, Front propagation in periodic excitable media, Com- munications on Pure and Applied Mathematics, LV, 949-1032, 2002

work page 2002
[2]

Berestycki, F

H. Berestycki, F. Hamel and N. Nadirashvili, The speed of propagation for KPP type problems. I-Periodic framework, J. Eur. Math. Soc., 7: 173-213, 2005

work page 2005
[3]

Berestycki, F

H. Berestycki, F. Hamel and L. Roques, Analysis of the periodically fragmented environment model: I-Species persistence, J. Math. Biol. 51: 75-113, 2005

work page 2005
[4]

Berestycki, F

H. Berestycki, F. Hamel and L. Roques, Analysis of the periodically fragmented environment model: I I -Biological invasions and pulsating traveling fronts, J. Math. Pures Appl. 84: 1101-1146, 2005

work page 2005
[6]

Fisher, The wave of advance of advantageous genes , Ann

R.A. Fisher, The wave of advance of advantageous genes , Ann. Eugenics 7, 335?369, 1937

work page 1937
[7]

Ito, Analysis of the minimal traveling wave speed via the methods of Young measures, SIAM J

R. Ito, Analysis of the minimal traveling wave speed via the methods of Young measures, SIAM J. Math. Anal., 50 (2018), pp. 3478-3534

work page 2018
[8]

Kinezaki, K

N. Kinezaki, K. Kawasaki, F. Takasu and N. Shigesada, Spatial dynamics of invasion in sinusoidally varying environments , Popul. Ecol., 48: 263-270, 2010. 14

work page 2010

Show all 17 references

[9]

Kolmogorov, I

A. Kolmogorov, I. Petrovsky, N. Piskunov, Etude de lf´ equation de la diﬀusion avec croissance de la quantit´ e de mati` ere et son application ` a un probl` eme biologique, Bjul. Moskowskogo Gos. Univ. Ser. Internat. Sec. A1: 1-26, 1937

1937
[10]

Liang, X

X. Liang, X. Lin and H. Matano, A variational problem associated with the min- imal speed of travelling wave for spatially periodic reaction-diﬀusion equation , Trans. Amer. Math. Soc., 362: 5605-5633, 2010

2010
[11]

Liang and H

X. Liang and H. Matano, Maximizing the spreading speed of KPP fronts in two dimensional stratiﬁed media, Pro. London Math. Soc., 109: 1137-1174, 2014

2014
[12]

R. Mori, D. Xiao, A variational problem associated with the minimal speed of traveling waves for spatially periodic KPP type equations , Proc. Lond. Math. Soc. 119, 654-680, 2019

2019
[13]

Nadin, The eﬀect of the Schwarz rearrangement on the periodic principal eigen- value of a nonsymmetric operator , SIAM J

N. Nadin, The eﬀect of the Schwarz rearrangement on the periodic principal eigen- value of a nonsymmetric operator , SIAM J. Math. Anal., 41:2388-2406, 2010

2010
[14]

Nadin, Some dependence results between the spreading speed and the coeﬃcients of the space-time periodic Fisher-KPP equation, European J

N. Nadin, Some dependence results between the spreading speed and the coeﬃcients of the space-time periodic Fisher-KPP equation, European J. Appl. Math., 22:169- 185, 2011

2011
[15]

Shigesada, K

N. Shigesada, K. Kawasaki and E. Teramoto, Traveling periodic waves in hetero- geneous environments, Theor. Population Biol., 30: 143-160, 1986

1986
[16]

J. G. Skellam Random dispersal in theoretical populations , Biometrika. 38: 196- 218, 1951

1951
[17]

H. F. Weinberger, On spreading speeds and traveling waves for growth and migra- tion models in a periodic habitat , J. Math. Biol. 45: 511-548, 2002. 15

2002

Pith tools

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

[1] [5]

ElSmaily, F

M. ElSmaily, F. Hamel and L. Roques, Homogenization and inﬂuence of frag- mentation in a biological invasion model , Disc. Cont. Dyn. Syst. A, 25: 321-342, 2009

work page 2009

[2] [1]

Berestycki and F

H. Berestycki and F. Hamel, Front propagation in periodic excitable media, Com- munications on Pure and Applied Mathematics, LV, 949-1032, 2002

work page 2002

[3] [2]

Berestycki, F

H. Berestycki, F. Hamel and N. Nadirashvili, The speed of propagation for KPP type problems. I-Periodic framework, J. Eur. Math. Soc., 7: 173-213, 2005

work page 2005

[4] [3]

Berestycki, F

H. Berestycki, F. Hamel and L. Roques, Analysis of the periodically fragmented environment model: I-Species persistence, J. Math. Biol. 51: 75-113, 2005

work page 2005

[5] [4]

Berestycki, F

H. Berestycki, F. Hamel and L. Roques, Analysis of the periodically fragmented environment model: I I -Biological invasions and pulsating traveling fronts, J. Math. Pures Appl. 84: 1101-1146, 2005

work page 2005

[6] [6]

Fisher, The wave of advance of advantageous genes , Ann

R.A. Fisher, The wave of advance of advantageous genes , Ann. Eugenics 7, 335?369, 1937

work page 1937

[7] [7]

Ito, Analysis of the minimal traveling wave speed via the methods of Young measures, SIAM J

R. Ito, Analysis of the minimal traveling wave speed via the methods of Young measures, SIAM J. Math. Anal., 50 (2018), pp. 3478-3534

work page 2018

[8] [8]

Kinezaki, K

N. Kinezaki, K. Kawasaki, F. Takasu and N. Shigesada, Spatial dynamics of invasion in sinusoidally varying environments , Popul. Ecol., 48: 263-270, 2010. 14

work page 2010