2104.01866
KAM below Cn
Jürgen Pöschel
correcthigh confidence
- Category
- Not specified
- Journal tier
- Specialist/Solid
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The paper proves a “Differentiable KAM Theorem” for rotational flows on T^n: if ω is Diophantine with exponent τ and the perturbation P is small in the block-L2 Besov-type norm ‖P‖_{τ+1,b}, then there exist a constant vector field Y and a C^1 diffeomorphism Ψ with Ψ^*(N+P−Y)=N; see the statement and the norm definition early in the paper and the iterative scheme in Sections 3–5 . The candidate solution reaches the same conclusion under the same smallness condition, but via a different route: a Littlewood–Paley/Nash–Moser KAM-by-blocks scheme directly in B^{τ+1}_{2,1}, solving the cohomological equation at each step and controlling commutators and remainders. The paper’s proof uses analytic smoothing, a Step Lemma based on a finite-dimensional fixed-point argument, and an iteration with analytic-width bookkeeping; the model uses Besov algebra/commutator estimates and flow pullbacks. Both approaches are logically coherent and compatible with the Diophantine loss τ and the need for one extra derivative to control the flow. Minor details the model leaves implicit (e.g., explicit constants, bounds ensuring the overlap between low-pass tail and sup-norm constraints, and rigorous justification of the Lie-series in the B^{r}_{2,1} setting) are standard and fillable using known Besov/Kato–Ponce machinery. Hence, both are correct, with different proofs.
Referee report (LaTeX)
\textbf{Recommendation:} minor revisions
\textbf{Journal Tier:} specialist/solid
\textbf{Justification:}
The paper achieves a notable advance in differentiable KAM by proving persistence below C\^n under a natural block-Besov smallness. The argument is clean, self-contained, and uses a well-structured analytic smoothing/Step Lemma/iteration thread. The model solution offers a consistent alternative proof via Besov machinery, reinforcing the main result from a different angle. Minor clarifications about constants, embeddings, and explicit identification of the norm with Besov B\^{r}\_{2,1} would improve readability and uptake.