2008.10171
Diffusion bound for the nonlinear Anderson model
Hongzi Cong, Yunfeng Shi
correctmedium confidence
- Category
- Not specified
- Journal tier
- Strong Field
- Processed
- Sep 28, 2025, 12:55 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The uploaded paper (Cong–Shi, arXiv:2008.10171v1, dated Aug 24–25, 2020) proves precisely the power-law diffusion bound D(t) = ∑j j^2|qj(t)|^2 ≤ t^κ for the 1D discrete nonlinear Anderson model with constant nonlinearity δ>0 and small ϵ+δ, removing the decaying λj assumption of Bourgain–Wang and resolving Bourgain’s 2006 problem (see the statement around Theorem 1.2 and remarks) . Their proof introduces a new tame “norm” (Definition 2.1) to run a finite-step Birkhoff normal form near barriers of width ∼ ln j0, obtaining a remainder estimate |||R̃||| ≤ j0^{-3/κ} and, consequently, a uniform-in-time differential inequality for the exterior mass that yields the desired t^κ bound (Sections 3 and 5) . The candidate solution asserts the result was likely open as of the cutoff and outlines only a finite-window barrier scheme; it misses the paper’s core tame-norm mechanism and almost-sure multiscale measure control that make the argument global in time. Hence the paper is correct on the main claim, while the model’s status assessment is incorrect.
Referee report (LaTeX)
\textbf{Recommendation:} minor revisions
\textbf{Journal Tier:} strong field
\textbf{Justification:}
The paper resolves Bourgain’s diffusion-bound problem in the nonlinear Anderson model without the previously essential decay of the nonlinearity, using a novel tame-norm–based barrier normal form. The result is important for the field of random dispersive equations and appears technically correct. Some expository improvements would aid accessibility, especially in the multiscale measure estimates and the final optimization step.