🔢 mathematics

On resonant energy sets for Hamiltonian systems with reflections

This paper investigates resonance in two uncoupled oscillators moving within a rectilinear polygon under elastic reflections, establishing a trichotomy for the size of resonant energy sets (empty, singleton, or open) and identifying a special class of potentials, $\mathcal{SP}$ , that allows for an abundance of resonant orbits, analogous to Bertrand's theorem.

Original authors: Krzysztof Frączek

Published 2026-07-28

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Krzysztof Frączek

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 universe where everything is made of tiny, invisible billiard balls bouncing around. In the world of physics, this is a common way to think about how particles move. Usually, we imagine these balls rolling on a flat table or bouncing inside a box. But what happens if the table isn't just a simple rectangle? What if it's a weird, jagged shape with sharp corners, like a maze made of straight lines? And what if the balls aren't just rolling freely, but are also being pulled by invisible springs that get stronger the further they stretch?

This is the playground of Hamiltonian systems, a branch of math and physics that studies how things move when they are governed by energy. In this specific story, we are looking at two separate "oscillators"—think of them as two independent springs, one moving left-right and the other moving up-down. They are trapped inside a polygon (a shape with straight sides) and bounce off the walls perfectly, like a mirror reflecting light. The big question scientists ask here is about resonance. Resonance is that magical feeling when you push a swing at just the right moment, and it goes higher and higher. In our bouncing ball world, resonance happens when the ball gets stuck in a perfect, repeating loop, hitting the same spots over and over again. If the system is "resonant," it's predictable and orderly. If it's not, the ball might bounce around chaotically, never repeating the same path twice. Understanding when these perfect loops happen helps us predict how complex systems behave, from the motion of planets to the vibration of atoms.

Now, enter the paper by Krzysztof Frączek. He decided to investigate exactly when these perfect, repeating loops (resonances) appear in our two-spring, bouncing-ball maze. He wasn't just looking for one lucky bounce; he wanted to know if there are entire "energy levels" (how fast the balls are moving) where resonance is guaranteed to happen for almost every possible starting position.

Here is the twist: Frączek discovered that resonance is actually quite picky. It doesn't happen just because you have a maze and some springs. It turns out that for resonance to be abundant (happening often and in big groups), the springs must be very special. Specifically, the "stiffness" of the springs (the mathematical shape of the potential energy) must be exactly quadratic—meaning they behave like perfect, ideal springs that follow a simple curve. If the springs are even slightly different (stiffer or softer in a weird way), the perfect loops almost entirely disappear.

The paper proves a fascinating "three-way split" (a trichotomy) regarding how many resonant energy levels exist:

Empty: There are no resonant levels at all. The balls bounce chaotically forever.
Singleton: There is exactly one specific energy level where resonance happens. It's a rare, isolated event.
Large (Open): There is a whole continuous range of energy levels where resonance is everywhere.

The most exciting part of the discovery is that the third option—the "Large" one—only happens if the springs are of a very specific, special type (which the author calls the SP class). However, having the special springs isn't enough on its own. If exactly one of the two springs belongs to this special class while the other is ordinary, the abundance of resonance vanishes completely, leaving you with no resonant levels at all.

Furthermore, even if you have the perfect springs (both of them in the special class), you still need a specific mathematical relationship between them. If the "stiffness" of the horizontal spring and the vertical spring are related by a simple fraction (a rational number), you get the abundance of loops. If that relationship is a messy, irrational number, the loops vanish.

This work is a bit like finding a secret rule for a video game. You might think that if you just build a complex maze, you'll get all sorts of cool, repeating patterns. But this paper proves that unless you build the maze with very specific, simple springs, tune both of them to be of that special type, and ensure their stiffness ratio is a simple fraction, the game will just be chaotic noise. It's a rigorous proof that nature is much more selective about when it allows for perfect, repeating order than we might have guessed. The authors didn't just guess this; they used advanced math involving "translation surfaces" (which is like unfolding the maze into a giant, flat sheet to see the paths more clearly) and deep analysis of how these paths behave to prove that these rules are absolute.

Title: On Resonant Energy Sets for Hamiltonian Systems with Reflections
Author: Krzysztof Frączek

Problem Statement

The paper investigates the dynamics of two uncoupled oscillators (one horizontal, one vertical) moving within a rectilinear polygon (a polygon with only vertical and horizontal sides) in the plane $\mathbb{R}^2$ . The system is governed by a Hamiltonian $H = \frac{1}{2}(p_1^2 + p_2^2) + V_1(q_1) + V_2(q_2)$ , where $V_1$ and $V_2$ are unimodal analytic potentials increasing away from zero. The particle undergoes elastic reflections at the polygon's boundary, with specific diffraction rules applied at concave corners.

The central problem is to characterize the set of resonant energy levels. An energy level $E > 0$ is defined as resonant if the pair $(E, \theta)$ is resonant for uncountably many $\theta \in (0, E)$ , where $\theta$ represents the energy partition between the two oscillators. A pair is resonant if the restricted flow on the invariant set $S_{E,\theta}$ possesses a resonant orbit (a periodic orbit, either regular or singular). The primary goal is to determine the "size" of the set of such resonant energy levels $\mathcal{E}(P, V_1, V_2)$ depending on the analytic properties of the potentials $V_1$ and $V_2$ .

Methodology

The paper employs a multi-faceted approach combining dynamical systems theory, complex analysis, and the geometry of translation surfaces:

Conjugacy to Billiards: The flow on the invariant set $S_{E,\theta}$ is conjugated to a billiard flow on a derived rectilinear polygon $P_{E,\theta}$ . This transformation utilizes scaled angle coordinates, mapping the oscillators' motion to linear motion with directions $\pm \pi/4, \pm 3\pi/4$ .
Translation Surfaces: The billiard flow on $P_{E,\theta}$ is further unfolded into a translation flow on a compact translation surface $(M, \omega)$ . Resonance in the billiard system corresponds to the existence of saddle connections (orbits joining singularities) on this surface.
Analytic Independence: The parameters of the polygon $P_{E,\theta}$ are expressed as integrals involving the potentials (functions denoted $a, \bar{a}, b, \bar{b}$ and their truncated versions $a_\xi, \bar{a}_\xi, \dots$ ). The paper analyzes the analytic properties of these functions, specifically their behavior at the endpoints of their domains. A key technical tool is the construction of differential operators $D_{\alpha, k}$ to demonstrate that linear combinations of these functions cannot vanish identically unless specific conditions on the potentials are met.
Geometric Quasi-Periodicity: In the case where resonant energy levels are abundant, the paper analyzes the functions $a + \bar{a}$ and $b + \bar{b}$ using the concept of geometric quasi-periodicity. This involves Fourier analysis of analytic functions satisfying $f(x+\tau) = \gamma f(x)$ . The authors prove that if such a function is positive and analytic, it must be constant, which imposes strict constraints on the potentials.

Key Contributions and Results

1. The Trichotomy of Resonant Energy Sets
The main result (Theorem 1.4) establishes a trichotomy for the set of resonant energy levels $\mathcal{E}(P, V_1, V_2)$ for any pair of unimodal analytic potentials and any rectilinear polygon:

Empty: The set is empty.
Singleton: The set contains exactly one element (which implies the degrees of the potentials at zero are both 2).
Large (Open and Non-empty): The set is an open interval.

2. Conditions for Abundance of Resonances
The "Large" case occurs if and only if:

Both potentials $V_1$ and $V_2$ belong to a special class denoted SP.
The ratio of the square roots of their second derivatives at zero, $\sqrt{V_1''(0)/V_2''(0)}$ , is rational.

3. Definition of the Special Class SP
The class SP consists of analytic potentials $V$ with $\deg(V, 0) = 2$ such that the inverse of the bi-analytic map $V_*$ (where $V_*^2 = V$ ) satisfies a specific symmetry condition: $V_*^{-1}(x) = cx + f(x^2)$ for some $c \neq 0$ and analytic $f$ .

If $V$ is even, $V \in SP$ implies $V$ is a quadratic function (harmonic oscillator).
If $V$ is not even, $V \in SP$ implies the oscillation frequency is constant (isochronous) for all energy levels.

4. Non-Resonance Results

Theorem 1.3: If at least one potential has a degree $m > 2$ at zero, or if both have degree 2 but are related by a scaling $V_1(x) = \omega V_2(\tau x)$ with an irrational ratio $\sqrt{V_1''(0)/V_2''(0)}$ , then the set of resonant energy levels is empty for any polygon.
If exactly one potential is in SP and the other is not, no resonant energy levels exist.
If neither is in SP, there can be at most one resonant energy level.

5. Analogy to Bertrand's Theorem
The paper draws a parallel between its findings and Bertrand's Theorem (1873). Just as Bertrand's Theorem characterizes centrally symmetric potentials (harmonic and Kepler) as the only ones admitting an abundance of periodic orbits, this work identifies the class SP as the unique family of potentials (in the context of uncoupled oscillators in polygons) that allow for an abundance of resonant energy levels.

Significance

The paper claims that detecting the "gap" in the size of the resonance energy set (i.e., that the set is either empty, a singleton, or open, but never a set of positive measure that is not open) is the most outstanding achievement of the work.

Novelty: The authors state that this phenomenon is new even in the simplest case where oscillators are not confined by polygonal barriers.
KAM Theory Motivation: The precise characterization of resonances is presented as a necessary prerequisite for applying KAM (Kolmogorov-Arnold-Moser) theory to perturbations of these quasi-integrable systems. Understanding the unperturbed resonance structure is essential for formulating the Diophantine conditions required for the persistence of invariant tori.
Generalization: The results extend previous studies on integrable systems with reflections (like the Boltzmann system) and near-integrable cases with even potentials to a much broader class of arbitrary rectilinear polygons and non-even analytic potentials.

The work does not propose new experimental applications but rather provides a rigorous theoretical framework for understanding the persistence of dynamics in perturbed Hamiltonian systems with reflections.

Problem Statement

Methodology

Key Contributions and Results

Significance

More like this