2208.10989
Detecting normally hyperbolic invariant tori in periodic non-autonomous differential equations
Pedro C.C.R. Pereira, Douglas D. Novaes, Murilo R. Cândido
correctmedium confidence
- Category
- Not specified
- Journal tier
- Strong Field
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The paper proves Theorem A: for ẋ = Σ_{i=1}^N ε^i F_i(t,x) + ε^{N+1} F̃(t,x,ε) with f_0 = ⋯ = f_{ℓ−1} = 0 and a guiding system ż = (1/T) f_ℓ(z) possessing an attracting hyperbolic limit cycle γ, there exists (for ε small) a T‑periodic solution γ_int and a normally hyperbolic attracting invariant torus that converges to γ × S^1; this is stated and proved via averaging, a moving frame, and a method-of-continuation (Theorem A and Theorem B) and a Brouwer fixed-point argument for γ_int . The candidate solution correctly applies higher-order averaging and a stroboscopic map expansion and uses normal hyperbolicity to persist an invariant closed curve for the map, and a degree argument to obtain a T‑periodic orbit. However, it then incorrectly asserts that M_ε := Γ_ε × S^1 is invariant for the continuous-time flow merely because it is invariant for the time‑T map. Invariance under the stroboscopic map does not by itself imply invariance under the full flow; the paper instead constructs the torus as the graph of a phase-dependent function h(σ,τ) (not a product set), ensuring flow invariance in the extended space . Hence the paper is correct, while the model’s proof contains a critical flaw in its final step.
Referee report (LaTeX)
\textbf{Recommendation:} minor revisions \textbf{Journal Tier:} strong field \textbf{Justification:} The main theorem cleanly links a hyperbolic attracting limit cycle in the guiding system to the existence of a normally hyperbolic invariant torus and an interior periodic orbit for the original periodic system, via a careful blend of higher-order averaging, a moving-frame reduction, and a continuation method. The argument is rigorous and well-motivated, with a useful application to jerk equations. Minor expository enhancements could further clarify the construction of the phase-dependent invariant graph and its relation to stroboscopic maps, but the results appear correct and of solid interest to the dynamics community.