REVIEW 1 major objections 5 minor 14 references
A Green Function Approach to Smooth Nonautonomous Topological Equivalence with Unbounded Nonlinearities under $(\mu,\nu)$--Dichotomies
T0 review · 1 major / 5 minor · reviewed 2026-08-27 · deepseek-v4-flash
Pith's one-line read The paper establishes that a nonautonomous linear system and its quasilinear perturbation are continuously, and under extra assumptions smoothly, topologically equivalent even when the nonlinear term is unbounded in the state variable.
desk verdict A solid, careful extension of Palmer-type linearization to (mu,nu)-dichotomies with unbounded nonlinearities; the proof is coherent and the examples are valid, but the key hypotheses (P6, R2) are not checkable from A and f alone. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the Green function $G(t,s)=T(t,s)P(s)$ for $t\ge s$ and $G(t,s)=-T(t,s)Q(s)$ for $t<s$, built from the invariant projections of the $(\mu,\nu)$-dichotomy, together with the Green-operator bound $C=\sup_{t\ge 0}\int_0^{+\infty}\|G(t,s)\|\,ds<\infty$. This kernel converts bounded solutions of the nonhomogeneous equation into fixed points of the contraction operator $\Gamma(\tau,\xi)\varphi(t)=\int_0^{+\infty}G(t,s)f(s,x(s,\tau,\xi)+\varphi(s))\,ds$, whose contraction constant is $q=C\gamma$. The novelty lies in assumptions (P5), (P6), and (R2): they require locally uniform $L^1$ majorants for $\|G(t,s)\|$ multiplied by the perturbation along linear, nonlinear, and variational trajectories, and it is these majorants that let limits pass through the integrals and allow differentiation under the integral sign. Condition (R3), a first-order smallness bound $\vartheta<1$ on the Green integral of the first variational derivative, guarantees that $D_\eta G(t,\eta)$ stays within an invertible neighborhood of the identity.
What would settle it
Compute $D_\eta G(t,\eta)=I-\int_0^{+\infty}G(t,s)D_yf(s,y(s,t,\eta))Y_1(s;t,\eta)\,ds$ for the diagonal example with $a=b=\kappa=1$, $\varepsilon=1/16$, and $\beta=2$, and then evaluate $D_\xi H(t,\xi)=(D_\eta G(t,H(t,\xi)))^{-1}$ near the diagonal $s=t$. If this derivative is singular or discontinuous for some $(t,\xi)$ while all assumptions (P1)-(P7) and (R1)-(R3) hold, the $C^r$ conclusion of Theorem 4.5 would be false; if it is invertible and continuous, the claimed smooth equivalence is verified in a concrete case.
Extended reading notes
Core claim
The central claim is Theorem 4.5: if the linear system (2.1) admits a $(\mu,\nu)$-dichotomy and assumptions (P1)-(P7) and (R1)-(R3) hold, then systems (2.1) and (2.2) are $C^r$-continuously topologically equivalent on $\mathbb R^+$. The equivalence is given explicitly by $H(t,\xi)=\xi+z^*(t;(t,\xi))$ and its inverse $G(t,\eta)=\eta+w^*(t;(t,\eta))$, where $z^*$ is the unique fixed point of the Green operator $\Gamma(\tau,\xi)$ acting on bounded continuous functions, and $w^*$ is the corresponding backward Green integral along the nonlinear solution. The homeomorphism sends every solution of the linear system to a solution of the nonlinear system and vice versa, is continuous in $(t,\xi)$, and grows unbounded in space. The $C^r$ conclusion follows by differentiating under the Green integral using locally uniform majorants for the variational derivatives $F_j(s;t,\eta)$, and by condition (R3), which gives $\|D_\eta G(t,\eta)-I\|\le\vartheta<1$ and hence invertibility of the derivative of the inverse map; differentiating the identity $G(t,H(t,\xi))=\xi$ recursively yields the continuity of all space derivatives of $H$.
Load-bearing premise
The whole construction depends on the assumption that the perturbation, evaluated along every nonlinear solution and every variational derivative, is locally uniformly integrable against the Green kernel: for each compact set of initial data there must be one $L^1$ majorant controlling all these compositions; this hypothesis involves the very solutions the theorem is meant to control, so it cannot be checked from $A$ and $f$ alone in general.
Editorial extensions
If this is right
- The continuous topological equivalence holds for unbounded perturbations under any $(\mu,\nu)$-dichotomy, not only for uniform contractions, so the result covers systems with both stable and unstable directions on the half-line.
- Smoothness of the equivalence is controlled separately: the topological homeomorphism needs only $q=C\gamma<1$, while the $C^r$ conclusion needs the variational Green-integrability (R2) and the invertibility condition $\vartheta<1$; one can hold without the other.
- For the worked examples, the abstract assumptions reduce to explicit inequalities, such as $2\varepsilon(1/a+1/b)<1$, $2\varepsilon e^{2\varepsilon/\beta}/\beta<1$, and $\beta>(r-1)b$, so the theorem gives checkable sufficient conditions for concrete systems.
- Because the Green-integrability conditions allow nonlinear solutions and their derivatives to grow, systems whose nonlinear solutions are unbounded in time can still have smooth linearizations, as long as the Green kernel compensates the growth.
Reading between the lines
- The same fixed-point-plus-Green-majorant scheme likely adapts to discrete-time and impulsive systems, replacing the Green integrals by Green sums; the paper does not state this, but nothing in the argument depends essentially on continuous time.
- A natural testable extension is to replace the global Lipschitz constant $\gamma$ in $q=C\gamma<1$ with a locally uniform Lipschitz bound along the relevant trajectories; if that weaker condition suffices, the theorem would cover perturbations whose growth is not globally Lipschitz.
- The strict inequality $\beta>(r-1)b$ in the examples suggests that the order of differentiability is limited by how fast the perturbation decays relative to the unstable growth; this threshold is likely necessary, not an artifact of the proof, though the paper does not prove necessity.
- One could probe whether condition (R3) is truly needed for all points or only to rule out singularities of $D_\eta G$; the paper notes a weaker determinant version, so verifying the determinant directly might yield $C^r$ equivalence under fewer hypotheses.
Formalized claims in Lean
-
Claim #1: The central claim is Theorem 4.5: if the linear system (2.1) admits a $(\mu,\nu)$-dichotomy and assumptions (P1)-(P7) and (R1)-(R3) hold, then systems (2.1) and (2.2) are $C^r$-continuously topologically equivalent on $\mathbb R^+$. The equivalence is given explicitly by $H(t,\xi)=\xi+z^*(t;(t,\xi))$ and its inverse $G(t,\eta)=\eta+w^*(t;(t,\eta))$, where $z^*$ is the unique fixed point of the Gre
/-- @claim 1 The central claim is Theorem 4.5: if the linear system (2.1) admits a $(\mu,\nu)$-dichotomy and assumptions (P1)-(P7) and (R1)-(R3) hold, then systems (2.1) and (2.2) are $C^r$-continuously topologically equivalent on $\mathbb R^+$. The equivalence is given explicitly by $H(t,\xi)=\xi+z^*(t;(t,\xi))$ and its inverse $G(t,\eta)=\eta+w^*(t;(t,\eta))$, where $z^*$ is the unique fixed point of the Gre -/ def central_claim : Prop :=
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the C^r topological equivalence between a nonautonomous linear system x'=A(t)x on R+ and a quasilinear perturbation y'=A(t)y+f(t,y), where f is globally Lipschitz in y but not necessarily bounded. The linear system is assumed to admit a (mu,nu)-dichotomy, and the perturbation is controlled by a growth function omega(t,r) through locally uniform Green-integrability conditions (P5)-(P6) along linear and nonlinear solutions. The authors construct Palmer-type maps H and G via fixed points of an integral operator and prove, under a smallness condition q=C gamma <1, that the systems are continuously topologically equivalent (Theorem 3.2). Under additional Green-integrability assumptions on the variational equations (R1)-(R3), they prove the equivalence is C^r (Theorem 4.5). Two classes of examples with diagonal linear systems and exponentially (or exponentially-polynomial) decaying unbounded perturbations are provided to illustrate the hypotheses.
Significance. If correct, this result extends the Green-function approach to smooth linearization to unbounded nonlinearities in the presence of a general (mu,nu)-dichotomy, without requiring boundedness of the nonlinear solutions. The proof is carefully structured: the fixed-point construction is standard, the consistency identities are correct, the dominated-convergence arguments are valid under the stated hypotheses, and the derivative-inversion step via R3 is clean. The examples are not curve-fitted; they verify all hypotheses explicitly via Gronwall estimates and cancellation properties of the Green function. The main weakness is that the key hypotheses P6 and R2 are formulated in terms of the unknown nonlinear flow and its derivatives, which severely limits direct verification from the data A and f in concrete applications.
major comments (1)
- [Section 2.3 (P6) and Section 4.1 (R2)] Assumptions (P6) and (R2) are not checkable from A and f alone: they require locally uniform Green-integrability of the perturbation along all nonlinear solutions y(s,tau,eta) and along all variational derivatives F_j(s;t,eta), i.e., along the very objects the theorem is supposed to control. These hypotheses are load-bearing: without them the fixed point w*, the dominated convergence step in the proof of Theorem 3.2, and the differentiation under the integral sign in Theorem 4.5 all fail. Remark 4.2's 'directly verifiable' sufficient condition still requires uniform bounds on the unknown Y_m(s;t,eta), so it does not resolve the checkability problem. The authors should either provide a general sufficient condition expressed directly in terms of A, f, and the dichotomy data, or explicitly discuss the self-referential nature of these hypotheses and clarify that the theorem should be read as a conditional statement requiring a priori control of the nonlinear flow. The examples are valid, but they do not remedy the general limitation.
minor comments (5)
- [Section 3, Definition 3.1 and (3.7)] The symbol G is used both for the Green function G(t,s) and for the inverse map G(t,eta) in Definition 3.1 and equation (3.7). This notational clash makes the proof of Theorem 3.2 harder to follow; the inverse map should be denoted differently (for example, script G or Phi).
- [The proof of Theorem 3.2] The paragraph 'Hence, -w is a fixed point of Gamma(t,G(t,eta))...' appears twice; the second occurrence is redundant and should be removed or merged with the first.
- [The proof of Theorem 4.5] The convergence F_j(s;t_m,eta_m) -> F_j(s;t,eta) for fixed s is asserted without proof; it follows from the standard theorem on C^r dependence of solutions on initial data and parameters, but the authors should cite or briefly justify this fact.
- [Section 5] Both examples have nu(t)=1, so they are uniform dichotomies with possibly non-exponential growth rates. Since the paper's framework is designed for (mu,nu)-dichotomies with nu possibly growing, an example with nu not identically 1 would strengthen the claim that the hypotheses are tractable in the genuinely nonuniform setting.
- [Remark 2.4] The remark correctly emphasizes that pointwise finiteness of the Green integrals is not enough for dominated convergence, but the wording could also note that the locally uniform majorants in (P5)-(P6) are uniform in the initial data in compact sets, not only in (t,eta), since this uniformity is what the subsequent proofs use.
Circularity Check
No significant circularity: H and G are constructed from stated contraction and Green-integrability assumptions, with all properties verified directly; the self-citation [4] is contextual, not load-bearing.
full rationale
The derivation is self-contained in the sense relevant to circularity: the conclusion (R+-continuously and then C^r-continuously topologically equivalent) is not an input to any hypothesis, and no 'prediction' is obtained by refitting a parameter. H and G are explicitly defined via the unique fixed point of the contraction Γ(τ,ξ) and the integral w*(t;(τ,η)); well-definedness uses (P4)-(P6), the inverse property is verified by direct fixed-point identities and the variation-of-constants formula, and continuity follows from dominated convergence with the locally uniform majorants. The C^r part differentiates under the integral using (R1)-(R2); invertibility of D_ηG is a consequence of (R3), a stated first-order smallness condition, not a renaming of the conclusion. The only self-citation, [4] by Castañeda and Torres, is used as contextual background about the program being continued; it is not invoked as a proof ingredient, a uniqueness theorem, or a source of the main estimates. The strong hypotheses (P6) and (R2) quantify over nonlinear solutions and their variational derivatives, so they can be hard to verify in general, and the paper verifies them in the examples via Gronwall estimates; this is an applicability limitation, not circularity, because these hypotheses are stated rather than derived from the equivalence they imply.
Assumptions & free parameters
assumptions (14)
- standard math Standard ODE existence, uniqueness, and continuous dependence of solutions (Coddington-Levinson [5]).
- standard math Gronwall inequality and matrix exponential estimates (Proposition 2.5).
- standard math Dominated convergence theorem and differentiation under the integral sign.
- standard math Inverse function theorem for finite-dimensional maps.
- domain assumption P1: linear system admits a (mu,nu)-dichotomy with invariant projections and constants M, lambda.
- domain assumption P2: f is continuous and globally Lipschitz in y with constant gamma.
- domain assumption P3: there exists a growth bound omega(t,r), nondecreasing in r, for f.
- domain assumption P4: C = sup_{t>=0} integral ||G(t,s)|| ds is finite.
- ad hoc to paper P5: locally uniform Green-integrability of omega along all linear solutions over compact initial data.
- ad hoc to paper P6: locally uniform Green-integrability of omega along all nonlinear solutions over compact initial data.
- domain assumption P7: q = C gamma < 1.
- domain assumption R1: f(t,.) is C^r with continuous derivatives in (t,y).
- ad hoc to paper R2: Green-integrable majorants for the variational terms F_j along nonlinear solutions.
- ad hoc to paper R3: sup over (t,eta) of integral ||G(t,s) D_y f(s,y) Y_1(s;t,eta)|| ds <= theta < 1.
Cite this review
Pith. "Pith review of A Green Function Approach to Smooth Nonautonomous Topological Equivalence with Unbounded Nonlinearities under $(\mu,\nu)$--Dichotomies." pith.science (2026). https://pith.science/paper/JODQG7YQ
@misc{pith2026260814715,
author = {Pith},
title = {Pith review of: A Green Function Approach to Smooth Nonautonomous Topological Equivalence with Unbounded Nonlinearities under $(\mu,\nu)$--Dichotomies},
year = {2026},
howpublished = {\url{https://pith.science/paper/JODQG7YQ}},
note = {Machine review of arXiv:2608.14715}
}
abstract
We study the smooth topological equivalence between a nonautonomous linear system on the positive half-line and a quasilinear perturbation whose nonlinear part is not assumed to be globally bounded with respect to the state variable. The linear equation is assumed to admit a $(\mu,\nu)$--dichotomy, and the construction is carried out through the Green operator associated with this dichotomy. The usual global boundedness of the perturbation is replaced by locally uniform Green-integrability conditions along the relevant linear and nonlinear solutions. Under a smallness condition involving the global Lipschitz constant of the perturbation and the Green operator, we first construct Palmer-type maps which give a continuous topological equivalence on $\mathbb R^+$. We then impose Green-integrability conditions on the successive variational equations and a further first-order smallness condition which guarantees the invertibility of the derivative of the inverse map. Under these assumptions, the equivalence is of class $C^r$. We also provide a class of examples for which the perturbation is unbounded with respect to the space variable and all the hypotheses can be verified directly.
Reference graph
Works this paper leans on
-
[4]
´A. Casta˜ neda and F. Torres, A nonautonomousC r–topological equivalence involving contractions and un- bounded nonlinearities. Math. Nachr.298(2025), 3893–3907. 14
work page 2025
-
[6]
D. Dragiˇ cevi´ c, W. Zhang, and W. Zhang, Smooth linearization of nonautonomous differential equations with a nonuniform dichotomy. Proc. Lond. Math. Soc. 121 (2020), 32–50
work page 2020
-
[1]
´A. Casta˜ neda, P. Monz´ on, and G. Robledo, Nonuniform contractions and density stability results via a smooth topological equivalence. Dyn. Syst. 38 (2023), 179–196
work page 2023
-
[2]
´A. Casta˜ neda, P. Monz´ on, and G. Robledo, Smoothness of topological equivalence on the half–line for nonau- tonomous systems. Proc. Roy. Soc. Edinburgh Sect. A 150 (2020), 2484–2502
work page 2020
-
[3]
´A. Casta˜ neda and G. Robledo, Differentiability of Palmer’s linearization theorem and converse result for density functions. J. Differential Equations 259 (2015), 4634–4650
work page 2015
-
[5]
E. Coddington, N. Levinson.Theory of Ordinary Differential Equations(Mc Graw–Hill: New York, 1955)
work page 1955
-
[7]
Hartman, A lemma in the theory of structural stability of differential equations
P. Hartman, A lemma in the theory of structural stability of differential equations. Proc. Amer. Math. Soc. 11 (1960), 610–620
work page 1960
-
[8]
I. Huerta. Linearization of a nonautonomous unbounded system with nonuniform contraction: a spectral ap- proach. Discrete Contin. Dyn. Syst. 40 (2020), 5571–5590
work page 2020
Show all 14 references
-
[9]
Jara, Smoothness of classC 2 of nonautonomous linearization without spectral conditions
N. Jara, Smoothness of classC 2 of nonautonomous linearization without spectral conditions. J. Dynam. Dif- ferential Equations 36 (2024), 1759–1776
2024
-
[10]
Lin, Hartman’s linearization on nonautonomous unbounded system
F. Lin, Hartman’s linearization on nonautonomous unbounded system. Nonlinear Anal. 66 (2007), 38–50
2007
-
[11]
K. J. Palmer, A generalization of Hartman’s linearization theorem. J. Math. Anal. Appl. 41 (1973), 753–758
1973
-
[12]
Reinfelds, D
A. Reinfelds, D. ˇSteinberga. Dynamical equivalence of impulsive quasilinear equations, Tatra Mt. Math. Publ.63 (2015), 237–246
2015
-
[13]
Pugh, On a theorem of P
C. Pugh, On a theorem of P. Hartman. Amer. J. Math. 91 (1969), 363–367
1969
-
[14]
Shi, K.Q
J.L. Shi, K.Q. Xiong. On Hartman’s linearization theorem and Palmer’s linearization theorem, J. Math. Anal. Appl.192 (1995), 813–832. Universidad de Chile, Departamento de Matem´aticas. Casilla 653, Santiago, Chile Email address:fernanda.torres.t@ug.uchile.cl
1995
Reviewed August 27, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.