🌀 nonlinear sciences

Recurrence in two degrees of freedom Hamiltonian flows

This paper demonstrates that recurrence time entropy (RTE) is an effective diagnostic tool for characterizing weak chaos and stickiness in two-degree-of-freedom Hamiltonian flows, successfully distinguishing regular, chaotic, and sticky regions in the Hénon-Heiles system with results consistent with established methods like the largest Lyapunov exponent and SALI.

Original authors: Matheus Rolim Sales, Leonardo Costa de Souza, Iberê Luiz Caldas, Edson Denis Leonel, José Danilo Szezech Jr

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Matheus Rolim Sales, Leonardo Costa de Souza, Iberê Luiz Caldas, Edson Denis Leonel, José Danilo Szezech Jr

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 you are watching a pinball machine, but instead of a simple game, the board is a complex, shifting landscape of hills, valleys, and invisible walls. In physics, this is called a Hamiltonian system. Sometimes, the ball (or a particle) moves in a perfectly predictable, rhythmic pattern, like a clock ticking. Other times, it goes wild, bouncing around chaotically.

But the most interesting part happens in the middle. Sometimes, a chaotic ball gets "stuck" near a calm, rhythmic area. It looks like it's behaving normally for a while, but deep down, it's still chaotic. This is called "stickiness." It's like a dancer who usually spins wildly but occasionally gets caught in a slow, graceful waltz before breaking back into a frenzy.

The problem for scientists is that this "sticky" behavior is hard to spot. If you watch for a short time, you might think the dancer is actually graceful (regular) when they are just temporarily stuck.

The New Tool: The "Recurrence Time Entropy" (RTE)

The authors of this paper introduce a new way to measure what's happening, called Recurrence Time Entropy (RTE).

Think of a Recurrence Plot as a diary of the dancer's steps. Every time the dancer returns to a spot they've visited before, you make a mark.

Regular dancers (predictable) return to the same spots in a very orderly, repeating pattern. Their diary looks like neat, long lines.
Chaotic dancers (wild) return to spots randomly. Their diary looks like scattered, short dashes.
Sticky dancers (the tricky ones) switch between neat lines and scattered dashes.

The RTE is a score that measures how "messy" or "complex" this diary is.

Low RTE Score: The pattern is simple and ordered (Regular motion).
High RTE Score: The pattern is messy and random (Chaotic motion).
Medium RTE Score: The pattern is a mix (Sticky motion).

What They Did

The researchers tested this new "scorecard" on two famous physics models:

The Hénon–Heiles System: Imagine a star moving in a galaxy. It's a system with two degrees of freedom (moving in two directions). They mapped out the star's path and calculated the RTE score for every starting point.
- The Result: The RTE score perfectly matched what scientists already knew using more complicated math tools (like the Lyapunov exponent). It correctly identified the calm "islands" of order, the wild "seas" of chaos, and the "sticky" borders where the star gets trapped temporarily.
The Paradigm Hamiltonian: This is a system with one degree of freedom that is being pushed by a rhythmic force (like a pendulum being pushed by a wave). They used a "stroboscopic" view (taking snapshots at regular intervals) to apply the same RTE test.
- The Result: It worked here too! The RTE score successfully distinguished between the calm and the chaotic, proving this tool isn't just a fluke for one specific system.

The "Stickiness" Discovery

The most exciting part of their discovery was looking at a single chaotic path over time. They watched how the RTE score changed second by second.

High RTE Episodes: When the score was high, the particle was exploring the wild, chaotic part of the system. These episodes lasted for a random amount of time, following a standard "exponential" pattern (like how long you wait for a bus that comes randomly).
Low RTE Episodes: When the score dropped, the particle had gotten "stuck" near a calm island. These episodes were different. They didn't follow a standard pattern; instead, they followed a "power law." This means that while most sticky traps are short, some last incredibly long, much longer than you would expect by chance. It's like getting stuck in a traffic jam that could last 5 minutes or 5 hours, with no predictable limit.

Why This Matters

Scientists have used other tools to find chaos, but many of those tools require knowing the exact equations of motion and doing heavy calculus (calculating how tiny changes grow).

The RTE is special because it only needs a record of where the particle was (a time series). It doesn't need the complex math equations.

They compared their RTE results with a famous, complex method called SALI (Smaller Alignment Index).
The Verdict: The simple RTE method gave the exact same results as the complex SALI method.

The Bottom Line

This paper shows that you can use a simple "messiness score" (RTE) based on how often a system returns to previous states to perfectly identify:

Orderly motion (Low score).
Wild chaos (High score).
The tricky "sticky" zones where chaos gets temporarily trapped (Medium score).

It's a simpler, more direct way to see the hidden structure in complex, chaotic systems, without needing to solve difficult differential equations. It confirms that "stickiness" is a real, measurable phenomenon where particles get trapped in order for surprisingly long times before escaping back into chaos.

Technical Summary: Recurrence in Two Degrees of Freedom Hamiltonian Flows

Problem Statement
Mixed Hamiltonian systems, characterized by a phase space containing both regular islands and chaotic seas, present a significant challenge in nonlinear dynamics. A primary difficulty is the phenomenon of "stickiness," where chaotic trajectories become temporarily trapped near regular structures (such as invariant tori or island boundaries) for extended, yet finite, periods. During these episodes, the motion appears regular over finite timescales, making it difficult to distinguish between regular, weakly chaotic, and strongly chaotic motion using standard finite-time diagnostics. Traditional measures, such as the Lyapunov exponent, can suffer from slow convergence due to these trapping events, while other tools may struggle to quantify the intermittent nature of weak chaos effectively.

Methodology
The authors investigate the efficacy of Recurrence Time Entropy (RTE) as a diagnostic tool for weak chaos and stickiness in Hamiltonian flows. The study employs two distinct models:

The Hénon–Heiles System: A paradigmatic autonomous Hamiltonian system with two degrees of freedom. The authors analyze trajectories by computing the RTE from successive crossings of a Poincaré Surface of Section (PSS).
The Paradigm Hamiltonian: A periodically driven, one-degree-of-freedom system (Escande's model). Here, the RTE is computed from a stroboscopic map.

The RTE is derived from the statistics of recurrence times, specifically calculated as the Shannon entropy of the distribution of white vertical lines (gaps between diagonal lines) in a Recurrence Plot (RP). The RP is constructed from the time series of intersection points on the PSS or stroboscopic map. The study utilizes a recurrence threshold $\varepsilon$ set to a fraction of the time series standard deviation ( $\varepsilon = \sigma/5$ ).

To validate the RTE, the authors perform several comparative analyses:

Correlation with LLE: Comparing RTE spatial distributions with the Largest Lyapunov Exponent (LLE).
Comparison with SALI: Assessing the proportion of chaotic trajectories identified by RTE against the Smaller Alignment Index (SALI), a variational method requiring the integration of deviation vectors.
Finite-Time Analysis: Analyzing RTE time series along single chaotic trajectories to identify "episodes" of low and high entropy.
Statistical Characterization: Computing the cumulative distributions of episode durations for low-RTE (trapping) and high-RTE (chaotic exploration) regimes.

Key Contributions and Results

Phase Space Characterization:
In the Hénon–Heiles system, the RTE successfully reproduces the global phase-space structures identified by the LLE. Regular islands correspond to low RTE values, chaotic regions to high RTE values, and the transition layers (sticky regions) exhibit intermediate RTE values. The RTE shows a strong positive correlation (Pearson coefficient > 0.94) with the LLE across various energy levels and threshold choices.
Detection of Stickiness and Intermittency:
The finite-time RTE analysis reveals strong fluctuations along individual chaotic trajectories. The authors identify distinct episodes:
- Low-RTE episodes: Associated with temporary trapping near regular islands. These correspond to time intervals where the recurrence-time distribution is less complex.
- High-RTE episodes: Associated with extended exploration of the chaotic sea, characterized by broader recurrence-time distributions.
  Projecting these episodes back onto the PSS confirms that low-RTE values are localized near regular islands and their hierarchical structures, while high-RTE values dominate the chaotic sea.
Statistical Signatures of Dynamics:
The study quantifies the duration ( $\tau$ ) of these episodes:
- Trapping (Low-RTE): The cumulative distribution of trapping durations decays algebraically ( $Q(\tau) \propto \tau^{-\gamma}$ ). The exponent $\gamma$ varies non-monotonically with energy, reflecting the complex reorganization of the phase-space hierarchy.
- Chaotic Exploration (High-RTE): The cumulative distribution of high-RTE episodes decays exponentially ( $Q(\tau) \propto e^{-\kappa\tau}$ ). The rate $\kappa$ decreases as energy increases, indicating that trajectories spend longer intervals in the strongly chaotic component as the chaotic sea expands.
Validation against Variational Methods:
The proportion of chaotic trajectories identified by the RTE (using a threshold of RTE > 2.5) aligns closely with the results obtained from the SALI (using a threshold of SALI < $10^{-10}$ ) across the energy range $E \in [1/10, 1/6]$ . Both indicators show a monotonic increase in the proportion of chaotic orbits as energy increases.
Generalizability:
The application of RTE to the periodically driven Paradigm Hamiltonian confirms that the method is not specific to the Hénon–Heiles model. It effectively distinguishes regular islands, chaotic seas, and sticky transition layers in time-periodic systems represented by stroboscopic maps.

Significance and Claims
The paper establishes the Recurrence Time Entropy (RTE) as an effective diagnostic tool for characterizing weak chaos and stickiness in Hamiltonian flows. The authors claim that RTE offers a distinct methodological advantage over variational indicators like the LLE and SALI: it relies solely on the trajectory time series and does not require the integration of variational equations or the computation of Jacobian matrices. This makes RTE particularly suitable for situations where only temporal data is available or where the equations of motion are difficult to differentiate.

The study concludes that RTE not only distinguishes between regular and chaotic regions but also captures the intermittent nature of sticky dynamics, recovering the expected algebraic statistics for trapping events and exponential statistics for chaotic exploration. By reproducing the phase-space structures seen in LLE and SALI analyses, the RTE is validated as a reliable, complementary tool for the study of mixed Hamiltonian dynamics.