Analyticity of Exponential Dirichlet Series and Applications to the Approximate Controllability of Parabolic Equations

Mohamed Ouzahra

REVIEW 2 major objections 5 minor 19 references

Analyticity of Exponential Dirichlet Series and Applications to the Approximate Controllability of Parabolic Equations

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper proves that a real exponential Dirichlet series is analytic under a simple summability condition and uses that to settle approximate controllability of the heat equation.

desk verdict A correct but low-novelty paper: solid analyticity proof for Dirichlet series, classical controllability applications, with a repairable WLOG gap in Corollary 2. read the letter →

arxiv 2506.07892 v1 pith:MUUZCFFI submitted 2025-06-09 math.OC math.AP

Mohamed Ouzahra This is my paper

classification math.OCmath.AP MSC 30B5035K0593B0593C20

keywords exponentialDirichletseriesanalyticityapproximatecontrollabilityheatequationmomentmethodlumpedcontroldistributedpowerremainder

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 a sufficient condition for a real exponential Dirichlet series $\varphi(t)=\sum_{j\ge 1}\alpha_j e^{-\lambda_j t}$ to be analytic on $(0,\infty)$: it is enough that the $\lambda_j$ are positive except for finitely many terms and that $\sum_j|\alpha_j|<\infty$, with a relaxed version allowing $\sum_j|\alpha_j|/|\lambda_j|^k<\infty$. It then spells out the Taylor coefficients $b_n=\sum_j \alpha_j e^{-\lambda_j\tau}(-\lambda_j)^n/n!$ at any $\tau>0$, together with a remainder estimate of order $|t-\tau|^n/(\tau^n\sqrt n)$. On the control-theory side, the paper uses this analyticity to prove that the heat equation on a bounded smooth domain is approximately controllable by controls supported on any non-null open set, and by a time-dependent lumped control exactly when every eigenfunction has nonzero integral over the control region. A sympathetic reader would care because it turns approximate controllability of a parabolic system into a checkable moment condition on the spectrum.

What carries the argument

The object that carries the argument is the exponential Dirichlet series $\varphi(t)=\sum_j\alpha_j e^{-\lambda_j t}$ and the observation that each term expands around $\tau$ as a power series in $t-\tau$. The key estimate is $e^{-\lambda_j\tau}\lambda_j^n\le(n/(e\tau))^n$ for every $j,n$, which after Stirling's bound makes the coefficient series $a_n=\sum_j|\alpha_j|e^{-\lambda_j\tau}\lambda_j^n/n!$ of order $O(1/(\tau^n\sqrt n))$; this guarantees the double series is absolutely summable on $(0,2\tau)$ and justifies interchanging the sums. The same coefficient bound controls the Taylor remainder. In the applications, the machinery is the moment-method equivalence: approximate controllability fails only if some nonzero state $y$ satisfies $B^*S^*(t)y=0$ for all $t\in[0,T]$, and expanding that condition in the eigenbasis produces an exponential Dirichlet series to which the uniqueness lemma applies.

What would settle it

On $\Omega=(0,1)$ with lumped control on $\omega=(0,\frac12)$, the paper predicts that global approximate controllability on $L^2(0,1)$ fails because $\int_0^{1/2}\sqrt2\sin(j\pi x)dx=0$ for every $j\equiv0\pmod4$. A numerical experiment attempting to steer the initial state $z_0=0$ to the target $\varphi_4(x)=\sqrt2\sin(4\pi x)$ within a prescribed tolerance, using $u(t)\in L^2(0,T)$, either would fail consistently, supporting the theorem, or would succeed, which would falsify the paper's criterion.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the moment method for parabolic control can be run on a standalone analyticity theorem: under $\sum_j|\alpha_j|<\infty$ with all but finitely many $\lambda_j>0$, the function $\varphi(t)=\sum_j\alpha_j e^{-\lambda_j t}$ is real-analytic on $(0,\infty)$, and at each $\tau>0$ it expands as $\varphi(t)=\sum_{n\ge 0}b_n(t-\tau)^n$ with $b_n=\sum_j\alpha_j e^{-\lambda_j\tau}(-\lambda_j)^n/n!$, the remainder after $n$ terms being $O(|t-\tau|^n/(\tau^n\sqrt n))$. A relaxed condition $\sum_j|\alpha_j|/|\lambda_j|^k<\infty$ gives the same analyticity for some $k\ge0$, and from it the paper derives a uniqueness lemma: if such a series vanishes on $[0,T]$, all coefficients $\alpha_j$ vanish. Applying the lemma to the eigenfunction expansion of the heat semigroup yields approximate controllability with distributed control on any non-null open set, and for lumped control $u(t)\mathbf{1}_\omega$ the necessary and sufficient condition $\int_\omega \varphi_j\ne0$ for every eigenfunction $\varphi_j$; in one dimension on $\omega=(a,b)\subset(0,1)$ this is equivalent to $a+b$ and $a-b$ being irrational.

Load-bearing premise

Two unproved premises carry the controllability conclusions: after the change of variable in the uniqueness lemma the shifted exponents still satisfy the summability condition, and a nonzero eigenfunction of the Dirichlet Laplacian cannot vanish on a set of positive measure; if either fails, the stated criteria would need to be re-examined.

Editorial extensions

If this is right

For any bounded smooth domain $\Omega$ and any non-empty open $\omega\subset\Omega$, the heat equation with control $u\in L^2(\Omega)$ acting as $\mathbf{1}_\omega u$ is approximately controllable on every interval $[0,T]$.
With a lumped time-dependent control, approximate controllability on $L^2(\Omega)$ holds exactly when $\int_\omega\varphi_j\ne0$ for every eigenfunction $\varphi_j$; on $\Omega=(0,1)$ and $\omega=(a,b)$, this is equivalent to $a+b\notin\mathbb{Q}$ and $a-b\notin\mathbb{Q}$.
When some eigenfunction has zero integral over $\omega$, global approximate controllability fails, but the system restricted to the orthogonal complement $V$ of the unreachable subspace is approximately controllable.
The explicit Taylor coefficients and the remainder estimate make the moment method quantitative: a truncated series approximates $\varphi$ with error $O(|t-\tau|^n/(\tau^n\sqrt n))$, so finite-dimensional control constructions inherit a rate.

Reading between the lines

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

The uniqueness lemma and the controllability criteria are stated for the Dirichlet Laplacian, but the same argument would apply to any self-adjoint generator with discrete simple spectrum and an eigenbasis for which the moment integrals can be computed; extending the criteria to other parabolic and damped-wave systems is a natural next step.
The remainder estimate suggests that the cost or the number of actuators needed to reach a given target within tolerance $\varepsilon$ should scale like a power of $1/\varepsilon$ governed by the smallest gap in the exponent sequence, although the paper does not compute such costs.
The proof of the uniqueness lemma shifts the exponents by $\lambda_1$; a complete version would verify that the shifted sequence still satisfies the summability hypothesis, and the distributed-control conclusion also depends on the unproved fact that a nonzero Laplacian eigenfunction cannot vanish on a set of positive measure.

Share X Bluesky LinkedIn Reddit HN

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 paper's real content is Theorem II.1: a self-contained proof that an exponential Dirichlet series with absolutely summable coefficients and eventually positive exponents is analytic on R+*, with explicit Taylor coefficients and a remainder bound. That proof is sound. The remainder estimate is a standard Taylor bound, but it is cleanly derived. The extension to H_k in Corollary 1 is indeed an elementary integration argument. The controllability applications are classical moment-method results, and the paper itself acknowledges recovering Lemma 3.14 of [4]. So the novelty is low, but the presentation is mostly careful and the main theorem is correct.

Where it gets soft: the proof of Corollary 2 has an unjustified 'without loss of generality' step. Shifting all λ_j by subtracting λ_1 sends the first exponent to zero, so the shifted sequence no longer satisfies H_k (which divides by λ_j^k). The statement 'By the previous corollary, φ is analytic' is not literally justified for the shifted sequence. This is a real rigor gap in the uniqueness lemma, but it is repairable: treat the constant term separately and apply Corollary 1 to the tail j≥2, whose exponents are positive. In the heat equation applications, the λ_j are −μ_j < 0 and the coefficients satisfy H_0 by Cauchy-Schwarz, so the shift is never needed; the controllability conclusions are not threatened. Still, the lemma as written needs a fix.

The other issues are minor: a sign error in Corollary 1's derivative (should be (−1)^k), the misstated definition of G in Section III(i) (φ(t) = ∑ e^{−λ_j t} ⟨y,φ_j⟩... missing the coefficient and sign), and λ_j = −(jπ)^2 given with a sign flip in the 1D lumped-control discussion. The title promises 'parabolic equations' but only the heat equation is treated.

The distributed-control result for any non-null open set and the lumped-control criterion ∫_ω φ_j ≠ 0 are already in the literature the paper cites, so a knowledgeable reader will not find new controllability theorems. But the paper does give a self-contained analyticity proof that could be pedagogically useful, and the applications are correct after the fixes.

Who is this for? Someone teaching the moment method from scratch, or a referee checking whether the analyticity lemma has a fully detailed proof. It deserves a serious referee: the main theorem is correct, the applications are valid, and the gaps are fixable. I would send it to peer review with a request for a revision that patches Corollary 2 and clears up the typos.

Referee Report

2 major / 5 minor

Summary. The paper studies real exponential Dirichlet series φ(t)=∑_{j≥1} α_j e^{-λ_j t} and proves (Theorem II.1) that under absolute summability of α_j and with all but finitely many λ_j positive, φ is analytic on (0,∞); it gives the Taylor coefficients b_n=∑_j α_j e^{-λ_j τ}(-λ_j)^n/n! around any τ>0 and the remainder estimate O(|t-τ|^n/(τ^n√n)). Corollary 1 relaxes the summability to (H_k), and Corollary 2 derives uniqueness of the coefficients from vanishing on an interval. The rest of the paper applies these results through the moment method to prove approximate controllability of the heat equation with distributed controls on any non-null open set and with lumped controls under the condition ∫_ω φ_j ≠0 for all j, including explicit 1D calculations.

Significance. If correct, the paper gives a self-contained proof of analyticity with explicit Taylor coefficients and a remainder rate, and a clean moment-method derivation of approximate controllability for the heat equation. The controllability conclusions themselves are consistent with the known literature; the main contribution is the explicit analyticity statement and remainder estimate. The writing is clear and the overall plan is sound. The proof of the key uniqueness lemma (Corollary 2) has a repairable gap, and there are several local errors; these should be fixed before publication.

major comments (2)

[II, Corollary 2] The reduction 'without loss of generality' in the proof of Corollary 2 is not justified as written. Multiplying (5) by e^{-λ_1 t} and replacing λ_j by λ_j-λ_1 produces a sequence whose first exponent is 0, so the shifted sequence is not contained in R*+ and Corollary 1 cannot be applied directly; the paper also does not verify that (H_k) holds for the shifted exponents. This gap is load-bearing because both controllability proofs in Section III invoke Corollary 2. The lemma is nevertheless true and the proof is repairable: since (λ_j) is strictly increasing, λ_j-λ_1 ≥ λ_2-λ_1 > 0 for j ≥ 2, and (H_k) implies ∑|α_j| < ∞, so the tail satisfies (H_k) with the shifted exponents; one should separate the constant j=1 term and apply Corollary 1 to the tail.
[III(A)] The distributed-control conclusion relies on the assertion that 1_ω φ_j ≠ 0 for every j whenever ω has non-null measure, i.e., on a unique continuation property for eigenfunctions of the Dirichlet Laplacian. The manuscript neither proves nor cites this property. Since it is essential for excluding nonzero elements of ker(G*), please add a precise statement and a reference to the unique continuation theorem used.

minor comments (5)

[II, Corollary 1] Equation (4) states that ϕ^(k)(t)=φ(t), but the k-th derivative of e^{-λ_j t} is (-λ_j)^k e^{-λ_j t}, so the correct relation is ϕ^(k)(t)=(-1)^k φ(t); the sign does not affect analyticity, but the displayed equality should be corrected.
[III(i)] The definition of the controllability operator is misstated: G u = ∫_0^t S(t-s)Bu(s)ds does not define a map U_T→H; it should read G u = ∫_0^T S(T-s)Bu(s)ds. The intended statement is clear from the later adjoint formula, but the displayed definition should be fixed.
[III(i), Theorem III.1] The statement of the approximate-controllability criterion is garbled: 'B*S*(t)y,∀t∈[0,T]⇒y=0' should read 'B*S*(t)y=0 for all t∈[0,T] implies y=0.'
[III(B), 1D case] In the one-dimensional lumped-control discussion, the text says λ_j=-(jπ)^2; since λ_j=-μ_j and μ_j=-(jπ)^2, the correct sign is λ_j=(jπ)^2. The printed sign is inconsistent with the notation e^{-λ_j t} used throughout.
[II, Theorem II.1] The reduction 'Without loss of generality, we can suppose that all the λ_j are positive' should be justified: the finitely many non-positive exponents give a finite sum of analytic functions, so one can subtract that finite sum and apply the argument to the tail with positive exponents.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the analyticity theorem is proved from first principles, and the controllability application relies on an external characterization plus the paper's own non-circular uniqueness corollary.

full rationale

The claimed derivation chain is self-contained and not circular. Theorem II.1 proves analyticity of φ(t)=Σα_j e^{-λ_j t} by expanding each exponential around τ and justifying the interchange with absolute convergence; the bound e^{-λ_jτ}λ_j^n ≤ (n/(eτ))^n together with Stirling gives a_n=O(1/(τ^n√n)), so the double series converges for t∈(0,2τ). Corollary 1 is a standard derivative reduction from H_k to H_0, and Corollary 2 is a uniqueness consequence obtained by analyticity and an induction on the coefficients. The heat-equation application does not fit any parameter to data and does not rename a fitted quantity as a prediction; it invokes the external controllability characterization (Theorem III.1, cited from [4]) and then uses Corollary 2 to force the coefficients α_j to vanish. There is no load-bearing self-citation chain and no ansatz smuggled in via citation. The softest point is the 'without loss of generality' shift in the proof of Corollary 2: subtracting λ_1 changes the exponent sequence and can put one exponent at 0, so the hypotheses of Corollary 1 are not literally verified as written; this is a rigor gap in an over-general lemma, not a circular step. In the applications H_0 holds directly with square-summable coefficients, so the shift is not needed for the controllability conclusion. The paper also asserts without proof that Dirichlet eigenfunctions do not vanish on any set of positive measure; that is an omitted external fact, not circularity. Overall, no step reduces by construction to its own input, so the circularity score is 0.

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

The paper introduces no new entities, numbers fitted to data, or ad hoc parameters. Its results rest on standard mathematical facts and on accepted control-theoretic characterizations. The main unstated assumptions are the uniqueness of eigenfunctions and the external controllability criterion.

assumptions (4)

domain assumption Approximate controllability is equivalent to ker(G*)={0} (Theorem III.1, cited from [4]).
The paper uses this characterization without proof to convert controllability into a condition on B*S*(t)y=0.
domain assumption Eigenfunctions of the Dirichlet Laplacian on a bounded domain cannot vanish on a set of positive measure.
Used to infer 1_ω φ_j ≠ 0 for any non-null open ω in the distributed control proof. Stated without proof.
standard math Standard analytic facts: Fubini for counting measure, Stirling's estimate, and the Taylor expansion of the exponential.
Used in the proof of Theorem II.1 to justify the double sum interchange and the asymptotic bound.
domain assumption The heat semigroup S(t) is self-adjoint and has a complete orthonormal eigenbasis with simple eigenvalues (assumed).
The spectral decomposition of the solution into eigenfunctions is used throughout Section III; simplicity is assumed and discussed only briefly.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Analyticity of Exponential Dirichlet Series and Applications to the Approximate Controllability of Parabolic Equations." pith.science (2026). https://pith.science/paper/MUUZCFFI

@misc{pith2026250607892,
  author       = {Pith},
  title        = {Pith review of: Analyticity of Exponential Dirichlet Series and Applications to the Approximate Controllability of Parabolic Equations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MUUZCFFI}},
  note         = {Machine review of arXiv:2506.07892}
}

read the original abstract

In this paper, we investigate the analyticity of a class of exponential Dirichlet series. We then explicitly determine the coefficients of their power series decomposition and provide an estimate for the remainder. As an application, we study the approximate controllability property of linear parabolic equations with locally distributed or lumped controls by employing the moment method, which relies on the exponential Dirichlet series associated with the spectrum of the system's operator.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

19 extracted references · 19 canonical work pages

[4]

F., & Pritchard, A

Curtain, R. F., & Pritchard, A. J. Infinite Dimensional Linear Systems Theory. Springer-Verlag, New York, 1978

work page 1978
[1]

Avdonin, S. A. E., Belishev, M. I., & Ivanov, S. A. (1992). Boundary control and a matrix inverse problem for the equation utt −u xx +V(x)u= 0. Sbornik: Mathematics, 72(2), 287-310

work page 1992
[2]

A., & Ivanov, S

Avdonin, S. A., & Ivanov, S. A. (1995). Families of exponentials: the method of moments in controllability problems for distributed parameter systems. Cambridge University Press

work page 1995
[3]

M., & Slemrod, M

Ball, J. M., & Slemrod, M. (1979). Feedback stabilization of distributed semi- linear control systems. Applied Mathematics and Optimization, 5(1), 169-179

work page 1979
[5]

Emanuilov, O. Y. Controllability conditions for the parabolic equations. Math- ematics of the USSR-Sbornik, 31.1 (1977): 141-153

work page 1977
[6]

El Jai, A., & Pritchard, A. J. (1987). Sensors and actuators in distributed systems. International Journal of Control, 46(4), 1139-1153

work page 1987
[7]

C., Zerrik, E., & Pritchard, A

El Jai, A., Simon, M. C., Zerrik, E., & Pritchard, A. J. (1995). Regional controllability of distributed parameter systems. International Journal of Control, 62(6), 1351-1365. 10

work page 1995
[8]

Fernández-Cara, E. (1997). Null controllability of the semilinear heat equation. ESAIM: Control, Optimisation and Calculus of Variations, 2, 87-103

work page 1997

Show all 19 references

[9]

and Imanuvilov, O

Fursikov A. and Imanuvilov, O. Yu. Controllability of evolution equations, Lecture Notes Series 32, Seoul National University, 1996

1996
[10]

H¨ormander, L. (2003). The analysis of linear partial differential operators. I. Classics in Mathematics

2003
[11]

Lebeau, G., & Robbiano, L. (1995). Contrôle exact de l’équation de la chaleur. Communications in Partial Differential Equations, 20(1-2), 335-356

1995
[12]

Pazy, A. (2012). Semigroups of linear operators and applications to partial differential equations (Vol. 44). Springer Science & Business Media

2012
[13]

Privat, Y., Trélat, E., & Zuazua, E. (2017). Actuator design for parabolic distributed parameter systems with the moment method. SIAM Journal on Control and Optimization, 55(2), 1128-1152

2017
[14]

Approximate controllability for the semilinear heat equation

Rosier, L. Approximate controllability for the semilinear heat equation. Jour- nal of Evolution Equations, 7.3 (2007): 549-569

2007
[15]

Russell, D. L. (1973). A unified boundary controllability theory for hyperbolic and parabolic partial differential equations. Studies in applied mathematics, 52(3), 189-211

1973
[16]

Russell, D. L. (1978). Controllability and stabilizability theory for linear partial differential equations: recent progress and open questions. Siam Review, 20(4), 639-739

1978
[17]

Triggiani, R. (1977). A note on the lack of exact controllability for mild solutions in Banach spaces. SIAM Journal on Control and Optimization, 15(3), 407-411

1977
[18]

Zuazua, E. (2000). Enrique Fernández-Cara. The cost of approximate control- lability for heat equations: the linear case. Advances in Differential Equations, 5(4-6), 465-514

2000
[19]

Controllability of partial differential equations and its applica- tions

Zuazua, E. Controllability of partial differential equations and its applica- tions. In Handbook of Differential Equations: Evolutionary Equations, Vol. 3, Elsevier, 2009, pp. 527-621. 11

2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [4]

F., & Pritchard, A

Curtain, R. F., & Pritchard, A. J. Infinite Dimensional Linear Systems Theory. Springer-Verlag, New York, 1978

work page 1978

[2] [1]

Avdonin, S. A. E., Belishev, M. I., & Ivanov, S. A. (1992). Boundary control and a matrix inverse problem for the equation utt −u xx +V(x)u= 0. Sbornik: Mathematics, 72(2), 287-310

work page 1992

[3] [2]

A., & Ivanov, S

Avdonin, S. A., & Ivanov, S. A. (1995). Families of exponentials: the method of moments in controllability problems for distributed parameter systems. Cambridge University Press

work page 1995

[4] [3]

M., & Slemrod, M

Ball, J. M., & Slemrod, M. (1979). Feedback stabilization of distributed semi- linear control systems. Applied Mathematics and Optimization, 5(1), 169-179

work page 1979

[5] [5]

Emanuilov, O. Y. Controllability conditions for the parabolic equations. Math- ematics of the USSR-Sbornik, 31.1 (1977): 141-153

work page 1977

[6] [6]

El Jai, A., & Pritchard, A. J. (1987). Sensors and actuators in distributed systems. International Journal of Control, 46(4), 1139-1153

work page 1987

[7] [7]

C., Zerrik, E., & Pritchard, A

El Jai, A., Simon, M. C., Zerrik, E., & Pritchard, A. J. (1995). Regional controllability of distributed parameter systems. International Journal of Control, 62(6), 1351-1365. 10

work page 1995

[8] [8]

Fernández-Cara, E. (1997). Null controllability of the semilinear heat equation. ESAIM: Control, Optimisation and Calculus of Variations, 2, 87-103

work page 1997