🔢 mathematics

Effective stability for Hamiltonian PDEs vanishing spectral gaps

This paper establishes effective stability for the space fractional Schrödinger equation in the regime $0 < \beta < 1/2$ by employing a high-low frequency decomposition to control near-resonance errors, thereby providing uniform stability time estimates for solutions across Gevrey, logarithmic ultra-differentiable, and finitely differentiable spaces.

Original authors: Bingqi Yu, Yong Li

Published 2026-07-15

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Bingqi Yu, Yong Li

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a vast, invisible ocean made of energy waves, swirling inside a mathematical box called a torus. This is the world of the fractional Schrödinger equation, a complex system that describes how particles behave when they are a bit "weird" and don't follow the usual rules of physics. In this ocean, energy usually tries to spread out, jumping from one wave to another, causing chaos. Scientists call this "energy transfer," and it's the kind of thing that makes systems unstable and unpredictable.

But what if we could prove that, under the right conditions, this energy stays put? What if we could show that the waves don't just scatter into chaos, but instead hold their shape for an incredibly long time? That is exactly what Bingqi Yu and Yong Li have done in their new paper. They are the lighthouse keepers of this mathematical ocean, showing us that for a specific, tricky type of wave system, stability is possible for a surprisingly long time.

The Tricky Part: The Vanishing Gaps

Usually, when scientists try to prove that a system is stable, they rely on a safety net called "spectral gaps." Think of these gaps like the rungs on a ladder. If the rungs are far apart, it's easy to tell one from the other, and you can build a sturdy ladder to climb up to stability.

However, the system Yu and Li studied is special. They are looking at a regime where the parameter $\beta$ is between 0 and 1/2. In this zone, the "rungs" of the ladder get closer and closer together as you go higher. Eventually, the gaps vanish completely. The frequencies of the waves become so crowded that they blur together. In the past, this "vanishing gap" problem was a nightmare for mathematicians because the standard tools they used to prove stability simply broke down. It was like trying to build a ladder where the rungs are melting into each other.

The Magic Trick: High-Low Frequency Decomposition

So, how did the authors fix a broken ladder? They didn't try to force the old tools to work. Instead, they invented a clever new strategy called high-low frequency decomposition.

Imagine you are trying to organize a massive library. The books on the bottom shelves (the low frequencies) are heavy, distinct, and easy to separate. The books on the top shelves (the high frequencies) are tiny, numerous, and packed so tightly they look like a blur.

The authors realized that while the low-frequency books are hard to organize because they are so close together, the high-frequency books are so small and light that they don't cause much trouble on their own. Their strategy was to:

Build a perfect, sturdy shelf for the low-frequency books. They used a mathematical technique called "Birkhoff normal form" to organize these main waves, eliminating the messy interactions that cause chaos.
Ignore the high-frequency blur for a moment. They treated the tiny, high-frequency waves as a "remainder." Because these waves are inherently small (if you start with a small disturbance), their messiness is weak.
Let the smallness do the heavy lifting. The key insight is that the errors caused by the messy high-frequency waves are so tiny that they get "absorbed" by the natural smallness of the solution itself. It's like a giant, slow-moving glacier (the stable low frequencies) swallowing up a few tiny pebbles (the high-frequency errors) without even noticing.

The Result: How Long is "Long"?

The paper proves that if you start with a very small amount of energy (represented by a tiny number $\epsilon$ ), the system will stay stable for a time that is sub-exponential.

To put this in perspective, imagine you drop a pebble into a calm pond.

In a chaotic system, the ripples would spread and crash into each other almost immediately.
In this new discovery, the ripples stay calm and organized for a time that is much longer than a polynomial (like $1/\epsilon^2$ ) but not quite as long as a pure exponential (like $e^{1/\epsilon}$ ).

The authors calculated the exact time limits for three different types of "smoothness" in the starting wave:

Gevrey Class (Super-smooth waves): The stability time is roughly $C_3 \exp\left(C_4 \frac{|\ln \epsilon|^{7/6}}{(\ln |\ln \epsilon|)^{1/6}}\right)$ . This is a very precise, long time, featuring a "logarithmic correction" that makes it slightly more accurate than previous guesses.
Logarithmic Ultra-differentiable (Very smooth waves): The time is $C_3 \exp\left(C_4 |\ln \epsilon|^{\frac{7q}{6q+1}}\right)$ .
Finite Differentiable (Rougher waves): The time is $C_3 \left(\frac{1}{\epsilon}\right)^{\frac{2C_4}{3} s^{1/7}}$ . This is a polynomial time, meaning it's shorter than the others but still significantly long.

What They Did NOT Find

It is important to note what this paper does not claim. The authors are not saying that all Hamiltonian systems are stable. They are not saying that energy never transfers. In fact, they explicitly focus on the regime $0 < \beta < 1/2$ . If the parameter $\beta$ were different (like the classic case where $\beta=1$ or $\beta=2$ ), the rules would be different, and the "vanishing gap" problem wouldn't exist in the same way. They are also not claiming that the system stays stable forever (infinite time); they are proving "effective stability," which means it stays stable for a very long, calculable time, but eventually, the tiny errors might add up.

The Bottom Line

Yu and Li have successfully navigated a mathematical minefield where the ground was supposed to be too shaky to stand on. By splitting the problem into "big, manageable waves" and "tiny, harmless ripples," they proved that even when the frequency gaps vanish, the system can remain calm for an incredibly long time. They didn't just guess; they provided rigorous mathematical proofs with explicit formulas for how long that stability lasts, depending on how smooth your starting wave is. It's a victory for order over chaos, showing that even in the most crowded, vanishingly small spaces, stability can hold its ground.

Technical Summary: Effective Stability for Hamiltonian PDEs with Vanishing Spectral Gaps

1. Problem Statement

This paper investigates the effective stability of the space fractional Schrödinger equation on a $d$ -dimensional torus $\mathbb{T}^d$ :
$i\partial_t\psi = (-\Delta+ m)^\beta\psi + p(|\psi|^2)\psi,$
where the parameter $\beta$ lies in the critical regime $0 < \beta < 1/2$ .

The central challenge addressed is the vanishing spectral gap. In standard Hamiltonian PDEs (e.g., classical NLS where $\omega_j \sim j^2$ ), the frequency separation $|\omega_i - \omega_j|$ is bounded away from zero, allowing for standard non-resonance conditions. However, for $0 < \beta < 1/2$ , the asymptotic frequencies $\omega_j = (|j|^2 + m)^\beta$ satisfy $\omega_{j+1} - \omega_j \to 0$ as $|j| \to \infty$ . This loss of frequency separation invalidates standard small divisor estimates used in KAM and Nekhoroshev theories, making the control of energy transfer (instability) significantly more difficult. The paper aims to establish explicit stability time scales for solutions with various regularities (Gevrey, logarithmic ultra-differentiable, and finitely differentiable) in this challenging regime.

2. Methodology

The authors employ a unified framework based on the Birkhoff normal form method, adapted to handle the vanishing spectral gaps through a high-low frequency decomposition.

2.1 High-Low Frequency Decomposition and Block Partitioning

To overcome the lack of global frequency separation, the phase space is partitioned into low-frequency modes ( $|j| \le N$ ) and high-frequency modes ( $|j| > N$ ). Crucially, the high-frequency modes are further partitioned into disjoint blocks $\Omega_\alpha$ with uniformly bounded "width."

Resonance Classification: Interactions are classified based on the number of high-frequency modes involved and whether they belong to the same block.
- Resonant interactions: Low-frequency exact resonances ( $R_0$ ) and high-frequency pairs within the same block with opposite signs ( $R_2$ ).
- Non-resonant interactions: Interactions involving modes from different blocks or single high-frequency modes ( $NR_{21}, NR_{22}, NR_1, NR_0$ ).
Small Divisor Estimates: By restricting non-resonant interactions to those between different blocks or specific configurations, the authors recover weak non-resonance conditions. They derive explicit lower bounds for the small divisors $|\sum \omega_{j_l}\sigma_l|$ that depend on the truncation scale $N$ and the block structure, specifically showing bounds of the form $\gamma N^{-C d^6}$ .

2.2 Iterative Normal Form Construction

The core of the analysis is an iterative lemma (Lemma 2) that constructs a sequence of canonical transformations to eliminate non-resonant terms order by order.

Homological Equation: At each step, the homological equation $\{H_0, G\} + P = Z$ is solved, where $Z$ is the resonant part and $G$ is the generating function. The solution $G$ is controlled by the small divisor estimates.
Error Control: The iteration generates remainder terms. The authors carefully track the dependence of these remainders on the truncation scale $N$ and the radius of the solution ball $r$ .
Balancing Remainders: A key technical step (Section 5) involves choosing the truncation scale $N$ as a function of the initial data size $r$ (and the regularity parameter) such that the "low-frequency" remainder (from the normal form iteration) and the "high-frequency" remainder (from the neglected tail) are of the same order of magnitude.

2.3 Stability Analysis

The stability time is derived by analyzing the evolution of the low-frequency part ( $w_<$ ) and the high-frequency part ( $w_>$ ) separately:

Low Frequencies: The growth is controlled by the smallness of the remainder terms, which decay exponentially with $N$ .
High Frequencies: The growth is controlled by the conservation of the $L^2$ norm and the inherent smallness of the initial high-frequency tail.
Synthesis: By combining these estimates, the authors derive explicit bounds on the time $T$ for which the solution remains within a small neighborhood of the initial state.

3. Key Contributions and Results

The paper establishes Nekhoroshev-type stability results for the fractional Schrödinger equation in the regime $0 < \beta < 1/2$ , providing explicit stability time estimates for three distinct regularity classes.

3.1 Gevrey Class ( $W^{s,\theta}_G$ )

For initial data in the Gevrey space with regularity $s > C_1$ and parameter $\theta > 0$ , the solution remains stable for a time:
$|t| \le C_3 \exp\left( C_4 \frac{|\ln \varepsilon|^{7/6}}{(\ln |\ln \varepsilon|)^{1/6}} \right)$
where $\varepsilon = \|\psi_0\|$ . This improves upon previous results of the form $\exp(C|\ln \varepsilon|^{1+a})$ by providing a more precise exponent with a logarithmic correction.

3.2 Logarithmic Ultra-differentiable Space ( $W^{s,q}_U$ )

For initial data in the logarithmic ultra-differentiable space with $q > 0$ , the stability time is:
$|t| \le C_3 \exp\left( C_4 |\ln \varepsilon|^{\frac{7q}{6q+1}} \right)$
This extends the known stability results for log-ultra-differentiable systems to the fractional Schrödinger equation with vanishing gaps.

3.3 Finitely Differentiable Space ( $W^s_F$ )

For initial data in the standard Sobolev space (finitely differentiable), the paper establishes polynomial stability:
$|t| \le C_3 \left( \frac{1}{\varepsilon} \right)^{\frac{2C_4}{3} s^{1/7}}$
This provides an explicit expression for the exponent $a(s)$ in the polynomial bound $(1/\varepsilon)^{a(s)}$ , refining previous qualitative results in the same regime.

4. Significance and Claims

The paper claims to be the first to systematically address the effective stability of Hamiltonian PDEs with asymptotic frequencies (vanishing spectral gaps) in the regime $0 < \beta < 1/2$ .

Overcoming the Gap: The primary contribution is the development of a block-partitioning strategy that allows for the recovery of weak non-resonance conditions despite the vanishing gaps. This enables the application of Birkhoff normal form techniques where standard methods fail.
Unified Framework: The authors demonstrate that a single analytical framework can yield stability estimates for Gevrey, log-ultra-differentiable, and finitely differentiable spaces, unifying the treatment of different regularity classes.
Explicit Time Scales: Unlike qualitative "almost global existence" results, this work provides explicit, quantitative time scales (sub-exponential and polynomial) dependent on the initial data size and regularity.
Refinement of Exponents: The paper claims to offer more precise estimates than previous works (e.g., [MC20]), particularly by deriving the specific exponents in the stability time formulas and including logarithmic corrections for the Gevrey case.

The results are presented as a rigorous extension of Nekhoroshev theory to a class of infinite-dimensional systems characterized by asymptotic frequency behavior, filling a gap in the literature regarding the long-time stability of fractional dispersive equations.